2006-02-24 04:48:15 +00:00
|
|
|
/*
|
|
|
|
** v_font.cpp
|
|
|
|
** Font management
|
|
|
|
**
|
|
|
|
**---------------------------------------------------------------------------
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
** Copyright 1998-2008 Randy Heit
|
2006-02-24 04:48:15 +00:00
|
|
|
** All rights reserved.
|
|
|
|
**
|
|
|
|
** Redistribution and use in source and binary forms, with or without
|
|
|
|
** modification, are permitted provided that the following conditions
|
|
|
|
** are met:
|
|
|
|
**
|
|
|
|
** 1. Redistributions of source code must retain the above copyright
|
|
|
|
** notice, this list of conditions and the following disclaimer.
|
|
|
|
** 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
** notice, this list of conditions and the following disclaimer in the
|
|
|
|
** documentation and/or other materials provided with the distribution.
|
|
|
|
** 3. The name of the author may not be used to endorse or promote products
|
|
|
|
** derived from this software without specific prior written permission.
|
|
|
|
**
|
|
|
|
** THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
|
|
|
|
** IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
|
|
|
|
** OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
|
|
|
|
** IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
|
|
|
|
** INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
|
|
|
|
** NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
|
|
|
** DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
|
|
|
** THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
|
|
** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
|
|
|
|
** THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
**---------------------------------------------------------------------------
|
|
|
|
**
|
|
|
|
*/
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
/* Special file formats handled here:
|
|
|
|
|
|
|
|
FON1 "console" fonts have the following header:
|
|
|
|
char Magic[4]; -- The characters "FON1"
|
|
|
|
uword CharWidth; -- Character cell width
|
|
|
|
uword CharHeight; -- Character cell height
|
|
|
|
|
|
|
|
The FON1 header is followed by RLE character data for all 256
|
|
|
|
8-bit ASCII characters.
|
|
|
|
|
|
|
|
|
|
|
|
FON2 "standard" fonts have the following header:
|
|
|
|
char Magic[4]; -- The characters "FON2"
|
|
|
|
uword FontHeight; -- Every character in a font has the same height
|
|
|
|
ubyte FirstChar; -- First character defined by this font.
|
|
|
|
ubyte LastChar; -- Last character definde by this font.
|
|
|
|
ubyte bConstantWidth;
|
|
|
|
ubyte ShadingType;
|
|
|
|
ubyte PaletteSize; -- size of palette in entries (not bytes!)
|
|
|
|
ubyte Flags;
|
|
|
|
|
|
|
|
There is presently only one flag for FON2:
|
|
|
|
FOF_WHOLEFONTKERNING 1 -- The Kerning field is present in the file
|
|
|
|
|
|
|
|
The FON2 header is followed by variable length data:
|
|
|
|
word Kerning;
|
|
|
|
-- only present if FOF_WHOLEFONTKERNING is set
|
|
|
|
|
|
|
|
ubyte Palette[PaletteSize+1][3];
|
|
|
|
-- The last entry is the delimiter color. The delimiter is not used
|
2008-11-27 17:43:36 +00:00
|
|
|
-- by the font but is used by imagetool when converting the font
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
-- back to an image. Color 0 is the transparent color and is also
|
|
|
|
-- used only for converting the font back to an image. The other
|
|
|
|
-- entries are all presorted in increasing order of brightness.
|
|
|
|
|
|
|
|
ubyte CharacterData[...];
|
|
|
|
-- RLE character data, in order
|
|
|
|
*/
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// HEADER FILES ------------------------------------------------------------
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
#include <stdlib.h>
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include <math.h>
|
|
|
|
#include <ctype.h>
|
|
|
|
|
|
|
|
#include "templates.h"
|
|
|
|
#include "doomtype.h"
|
|
|
|
#include "m_swap.h"
|
|
|
|
#include "v_font.h"
|
|
|
|
#include "v_video.h"
|
|
|
|
#include "w_wad.h"
|
|
|
|
#include "i_system.h"
|
|
|
|
#include "gi.h"
|
|
|
|
#include "cmdlib.h"
|
|
|
|
#include "sc_man.h"
|
2006-10-06 21:07:58 +00:00
|
|
|
#include "hu_stuff.h"
|
2011-07-06 10:55:04 +00:00
|
|
|
#include "farchive.h"
|
|
|
|
#include "textures/textures.h"
|
2011-07-06 08:50:15 +00:00
|
|
|
#include "r_data/r_translate.h"
|
2008-09-14 23:54:38 +00:00
|
|
|
#include "colormatcher.h"
|
2008-09-15 14:11:05 +00:00
|
|
|
#include "v_palette.h"
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// MACROS ------------------------------------------------------------------
|
|
|
|
|
2007-01-09 04:40:58 +00:00
|
|
|
#define DEFAULT_LOG_COLOR PalEntry(223,223,223)
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// TYPES -------------------------------------------------------------------
|
2008-09-15 14:11:05 +00:00
|
|
|
void RecordTextureColors (FTexture *pic, BYTE *colorsused);
|
2006-08-30 02:38:39 +00:00
|
|
|
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
// This structure is used by BuildTranslations() to hold color information.
|
|
|
|
struct TranslationParm
|
|
|
|
{
|
|
|
|
short RangeStart; // First level for this range
|
|
|
|
short RangeEnd; // Last level for this range
|
|
|
|
BYTE Start[3]; // Start color for this range
|
|
|
|
BYTE End[3]; // End color for this range
|
|
|
|
};
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
struct TranslationMap
|
|
|
|
{
|
|
|
|
FName Name;
|
|
|
|
int Number;
|
|
|
|
};
|
|
|
|
|
2008-09-15 14:11:05 +00:00
|
|
|
class FSingleLumpFont : public FFont
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
FSingleLumpFont (const char *fontname, int lump);
|
|
|
|
|
|
|
|
protected:
|
2012-08-06 10:25:51 +00:00
|
|
|
void CheckFON1Chars (double *luminosity);
|
2008-09-15 14:11:05 +00:00
|
|
|
void BuildTranslations2 ();
|
|
|
|
void FixupPalette (BYTE *identity, double *luminosity, const BYTE *palette,
|
|
|
|
bool rescale, PalEntry *out_palette);
|
2012-08-06 10:25:51 +00:00
|
|
|
void LoadTranslations ();
|
2008-09-15 14:11:05 +00:00
|
|
|
void LoadFON1 (int lump, const BYTE *data);
|
|
|
|
void LoadFON2 (int lump, const BYTE *data);
|
2010-02-26 05:34:30 +00:00
|
|
|
void LoadBMF (int lump, const BYTE *data);
|
2008-09-15 14:11:05 +00:00
|
|
|
void CreateFontFromPic (FTextureID picnum);
|
2010-02-26 05:34:30 +00:00
|
|
|
|
|
|
|
static int STACK_ARGS BMFCompare(const void *a, const void *b);
|
2012-08-06 10:25:51 +00:00
|
|
|
|
|
|
|
enum
|
|
|
|
{
|
|
|
|
FONT1,
|
|
|
|
FONT2,
|
|
|
|
BMFFONT
|
|
|
|
} FontType;
|
|
|
|
BYTE PaletteData[768];
|
|
|
|
bool RescalePalette;
|
2008-09-15 14:11:05 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
class FSinglePicFont : public FFont
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
FSinglePicFont(const char *picname);
|
|
|
|
|
|
|
|
// FFont interface
|
|
|
|
FTexture *GetChar (int code, int *const width) const;
|
|
|
|
int GetCharWidth (int code) const;
|
|
|
|
|
|
|
|
protected:
|
|
|
|
FTextureID PicNum;
|
|
|
|
};
|
|
|
|
|
2008-11-27 17:43:36 +00:00
|
|
|
// Essentially a normal multilump font but with an explicit list of character patches
|
|
|
|
class FSpecialFont : public FFont
|
|
|
|
{
|
|
|
|
public:
|
2012-08-06 10:25:51 +00:00
|
|
|
FSpecialFont (const char *name, int first, int count, FTexture **lumplist, const bool *notranslate, int lump);
|
|
|
|
|
|
|
|
void LoadTranslations();
|
|
|
|
|
|
|
|
protected:
|
|
|
|
bool notranslate[256];
|
2008-11-27 17:43:36 +00:00
|
|
|
};
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
// This is a font character that loads a texture and recolors it.
|
|
|
|
class FFontChar1 : public FTexture
|
|
|
|
{
|
|
|
|
public:
|
2012-08-06 10:25:51 +00:00
|
|
|
FFontChar1 (FTexture *sourcelump);
|
2006-02-24 04:48:15 +00:00
|
|
|
const BYTE *GetColumn (unsigned int column, const Span **spans_out);
|
|
|
|
const BYTE *GetPixels ();
|
2012-08-06 10:25:51 +00:00
|
|
|
void SetSourceRemap(const BYTE *sourceremap);
|
2006-02-24 04:48:15 +00:00
|
|
|
void Unload ();
|
|
|
|
~FFontChar1 ();
|
|
|
|
|
|
|
|
protected:
|
|
|
|
void MakeTexture ();
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
FTexture *BaseTexture;
|
2006-02-24 04:48:15 +00:00
|
|
|
BYTE *Pixels;
|
|
|
|
const BYTE *SourceRemap;
|
|
|
|
};
|
|
|
|
|
|
|
|
// This is a font character that reads RLE compressed data.
|
|
|
|
class FFontChar2 : public FTexture
|
|
|
|
{
|
|
|
|
public:
|
2012-08-06 10:25:51 +00:00
|
|
|
FFontChar2 (int sourcelump, int sourcepos, int width, int height, int leftofs=0, int topofs=0);
|
2006-02-24 04:48:15 +00:00
|
|
|
~FFontChar2 ();
|
|
|
|
|
|
|
|
const BYTE *GetColumn (unsigned int column, const Span **spans_out);
|
|
|
|
const BYTE *GetPixels ();
|
2012-08-06 10:25:51 +00:00
|
|
|
void SetSourceRemap(const BYTE *sourceremap);
|
2006-02-24 04:48:15 +00:00
|
|
|
void Unload ();
|
|
|
|
|
|
|
|
protected:
|
|
|
|
int SourceLump;
|
|
|
|
int SourcePos;
|
|
|
|
BYTE *Pixels;
|
|
|
|
Span **Spans;
|
2006-08-30 02:38:39 +00:00
|
|
|
const BYTE *SourceRemap;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
void MakeTexture ();
|
|
|
|
};
|
|
|
|
|
2006-08-31 01:30:22 +00:00
|
|
|
struct TempParmInfo
|
|
|
|
{
|
|
|
|
unsigned int StartParm[2];
|
|
|
|
unsigned int ParmLen[2];
|
|
|
|
int Index;
|
|
|
|
};
|
|
|
|
struct TempColorInfo
|
|
|
|
{
|
|
|
|
FName Name;
|
|
|
|
unsigned int ParmInfo;
|
2007-01-09 04:40:58 +00:00
|
|
|
PalEntry LogColor;
|
2006-08-31 01:30:22 +00:00
|
|
|
};
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// EXTERNAL FUNCTION PROTOTYPES --------------------------------------------
|
|
|
|
|
|
|
|
// PUBLIC FUNCTION PROTOTYPES ----------------------------------------------
|
|
|
|
|
|
|
|
// PRIVATE FUNCTION PROTOTYPES ---------------------------------------------
|
|
|
|
|
|
|
|
static int STACK_ARGS TranslationMapCompare (const void *a, const void *b);
|
|
|
|
|
|
|
|
// EXTERNAL DATA DECLARATIONS ----------------------------------------------
|
|
|
|
|
2009-02-20 00:53:25 +00:00
|
|
|
extern int PrintColors[];
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// PUBLIC DATA DEFINITIONS -------------------------------------------------
|
|
|
|
|
|
|
|
FFont *FFont::FirstFont = NULL;
|
2006-08-31 00:16:12 +00:00
|
|
|
int NumTextColors;
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
// PRIVATE DATA DEFINITIONS ------------------------------------------------
|
|
|
|
|
|
|
|
static TArray<TranslationParm> TranslationParms[2];
|
|
|
|
static TArray<TranslationMap> TranslationLookup;
|
2007-01-09 04:40:58 +00:00
|
|
|
static TArray<PalEntry> TranslationColors;
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
// CODE --------------------------------------------------------------------
|
|
|
|
|
2013-05-17 02:35:35 +00:00
|
|
|
static bool myislower(int code)
|
|
|
|
{
|
|
|
|
if (code >= 'a' && code <= 'z')
|
|
|
|
return true;
|
|
|
|
if (code != 0xF7 && code != 0xFF && (code & 0xE0) == 0xE0)
|
|
|
|
return true;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Returns a character without an accent mark.
|
|
|
|
// FIXME: Only valid for CP-1252; we should go Unicode at some point.
|
|
|
|
|
|
|
|
static int stripaccent(int code)
|
|
|
|
{
|
|
|
|
if (code < 0x8a)
|
|
|
|
return code;
|
|
|
|
if (code == 0x8a) // Latin capital letter S with caron
|
|
|
|
return 'S';
|
|
|
|
if (code == 0x8e) // Latin capital letter Z with caron
|
|
|
|
return 'Z';
|
|
|
|
if (code == 0x9a) // Latin small letter S with caron
|
|
|
|
return 's';
|
|
|
|
if (code == 0x9e) // Latin small letter Z with caron
|
|
|
|
return 'z';
|
|
|
|
if (code == 0x9f) // Latin capital letter Y with diaeresis
|
|
|
|
return 'Y';
|
|
|
|
if (code == 0xff) // Latin small letter Y with diaeresis
|
|
|
|
return 'y';
|
|
|
|
// Every other accented character has the high two bits set.
|
|
|
|
if ((code & 0xC0) == 0)
|
|
|
|
return code;
|
|
|
|
// Make lowercase characters uppercase so there are half as many tests.
|
|
|
|
int acode = code & 0xDF;
|
|
|
|
if (acode >= 0xC0 && acode <= 0xC5) // A with accents
|
|
|
|
return 'A' + (code & 0x20);
|
|
|
|
if (acode == 0xC7) // Cedilla
|
|
|
|
return 'C' + (acode & 0x20);
|
|
|
|
if (acode >= 0xC8 && acode <= 0xCB) // E with accents
|
|
|
|
return 'E' + (code & 0x20);
|
|
|
|
if (acode >= 0xCC && acode <= 0xCF) // I with accents
|
|
|
|
return 'I' + (code & 0x20);
|
|
|
|
if (acode == 0xD0) // Eth
|
|
|
|
return 'D' + (code & 0x20);
|
|
|
|
if (acode == 0xD1) // N with tilde
|
|
|
|
return 'N' + (code & 0x20);
|
2013-05-26 00:53:48 +00:00
|
|
|
if ((acode >= 0xD2 && acode <= 0xD6) || // O with accents
|
2013-05-17 02:35:35 +00:00
|
|
|
acode == 0xD8) // O with stroke
|
|
|
|
return 'O' + (code & 0x20);
|
|
|
|
if (acode >= 0xD9 && acode <= 0xDC) // U with accents
|
|
|
|
return 'U' + (code & 0x20);
|
|
|
|
if (acode == 0xDD) // Y with accute
|
|
|
|
return 'Y' + (code & 0x20);
|
|
|
|
if (acode == 0xDE) // Thorn
|
|
|
|
return 'P' + (code & 0x20); // well, it sort of looks like a 'P'
|
|
|
|
return code;
|
|
|
|
}
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
FFont *V_GetFont(const char *name)
|
2006-09-17 10:43:51 +00:00
|
|
|
{
|
|
|
|
FFont *font = FFont::FindFont (name);
|
|
|
|
if (font == NULL)
|
|
|
|
{
|
2007-12-23 20:59:06 +00:00
|
|
|
int lump = -1;
|
2008-03-29 22:59:41 +00:00
|
|
|
|
|
|
|
lump = Wads.CheckNumForFullName(name, true);
|
2007-12-23 20:59:06 +00:00
|
|
|
|
2006-09-17 10:43:51 +00:00
|
|
|
if (lump != -1)
|
|
|
|
{
|
2010-02-26 20:19:43 +00:00
|
|
|
uint32 head;
|
2006-09-17 10:43:51 +00:00
|
|
|
{
|
|
|
|
FWadLump lumpy = Wads.OpenLumpNum (lump);
|
2010-02-26 20:19:43 +00:00
|
|
|
lumpy.Read (&head, 4);
|
2006-09-17 10:43:51 +00:00
|
|
|
}
|
2010-02-26 20:19:43 +00:00
|
|
|
if ((head & MAKE_ID(255,255,255,0)) == MAKE_ID('F','O','N',0) ||
|
|
|
|
head == MAKE_ID(0xE1,0xE6,0xD5,0x1A))
|
2006-09-17 10:43:51 +00:00
|
|
|
{
|
|
|
|
font = new FSingleLumpFont (name, lump);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (font == NULL)
|
|
|
|
{
|
2008-06-15 18:36:26 +00:00
|
|
|
FTextureID picnum = TexMan.CheckForTexture (name, FTexture::TEX_Any);
|
|
|
|
if (picnum.isValid())
|
2006-09-17 10:43:51 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
font = new FSinglePicFont (name);
|
2006-09-17 10:43:51 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return font;
|
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// SerializeFFontPtr
|
|
|
|
//
|
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
FArchive &SerializeFFontPtr (FArchive &arc, FFont* &font)
|
|
|
|
{
|
|
|
|
if (arc.IsStoring ())
|
|
|
|
{
|
|
|
|
arc << font->Name;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
char *name = NULL;
|
|
|
|
|
|
|
|
arc << name;
|
2006-09-17 10:43:51 +00:00
|
|
|
font = V_GetFont(name);
|
2006-02-24 04:48:15 +00:00
|
|
|
if (font == NULL)
|
|
|
|
{
|
2006-09-17 10:43:51 +00:00
|
|
|
Printf ("Could not load font %s\n", name);
|
|
|
|
font = SmallFont;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
delete[] name;
|
|
|
|
}
|
|
|
|
return arc;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: FFont
|
|
|
|
//
|
|
|
|
// Loads a multi-texture font.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-07 08:11:56 +00:00
|
|
|
FFont::FFont (const char *name, const char *nametemplate, int first, int count, int start, int fdlump, int spacewidth)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
int i;
|
|
|
|
FTextureID lump;
|
2006-02-24 04:48:15 +00:00
|
|
|
char buffer[12];
|
2008-11-30 12:49:27 +00:00
|
|
|
FTexture **charlumps;
|
2006-02-24 04:48:15 +00:00
|
|
|
int maxyoffs;
|
2008-08-30 19:44:19 +00:00
|
|
|
bool doomtemplate = gameinfo.gametype & GAME_DoomChex ? strncmp (nametemplate, "STCFN", 5) == 0 : false;
|
2008-04-05 12:14:33 +00:00
|
|
|
bool stcfn121 = false;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
Lump = fdlump;
|
2006-02-24 04:48:15 +00:00
|
|
|
Chars = new CharData[count];
|
2008-11-30 12:49:27 +00:00
|
|
|
charlumps = new FTexture *[count];
|
2006-02-24 04:48:15 +00:00
|
|
|
PatchRemap = new BYTE[256];
|
|
|
|
FirstChar = first;
|
|
|
|
LastChar = first + count - 1;
|
|
|
|
FontHeight = 0;
|
|
|
|
GlobalKerning = false;
|
|
|
|
Name = copystring (name);
|
|
|
|
Next = FirstFont;
|
|
|
|
FirstFont = this;
|
2010-10-12 08:43:15 +00:00
|
|
|
Cursor = '_';
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
maxyoffs = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++)
|
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
charlumps[i] = NULL;
|
About a week's worth of changes here. As a heads-up, I wouldn't be
surprised if this doesn't build in Linux right now. The CMakeLists.txt
were checked with MinGW and NMake, but how they fair under Linux is an
unknown to me at this time.
- Converted most sprintf (and all wsprintf) calls to either mysnprintf or
FStrings, depending on the situation.
- Changed the strings in the wbstartstruct to be FStrings.
- Changed myvsnprintf() to output nothing if count is greater than INT_MAX.
This is so that I can use a series of mysnprintf() calls and advance the
pointer for each one. Once the pointer goes beyond the end of the buffer,
the count will go negative, but since it's an unsigned type it will be
seen as excessively huge instead. This should not be a problem, as there's
no reason for ZDoom to be using text buffers larger than 2 GB anywhere.
- Ripped out the disabled bit from FGameConfigFile::MigrateOldConfig().
- Changed CalcMapName() to return an FString instead of a pointer to a static
buffer.
- Changed startmap in d_main.cpp into an FString.
- Changed CheckWarpTransMap() to take an FString& as the first argument.
- Changed d_mapname in g_level.cpp into an FString.
- Changed DoSubstitution() in ct_chat.cpp to place the substitutions in an
FString.
- Fixed: The MAPINFO parser wrote into the string buffer to construct a map
name when given a Hexen map number. This was fine with the old scanner
code, but only a happy coincidence prevents it from crashing with the new
code
- Added the 'B' conversion specifier to StringFormat::VWorker() for printing
binary numbers.
- Added CMake support for building with MinGW, MSYS, and NMake. Linux support
is probably broken until I get around to booting into Linux again. Niceties
provided over the existing Makefiles they're replacing:
* All command-line builds can use the same build system, rather than having
a separate one for MinGW and another for Linux.
* Microsoft's NMake tool is supported as a target.
* Progress meters.
* Parallel makes work from a fresh checkout without needing to be primed
first with a single-threaded make.
* Porting to other architectures should be simplified, whenever that day
comes.
- Replaced the makewad tool with zipdir. This handles the dependency tracking
itself instead of generating an external makefile to do it, since I couldn't
figure out how to generate a makefile with an external tool and include it
with a CMake-generated makefile. Where makewad used a master list of files
to generate the package file, zipdir just zips the entire contents of one or
more directories.
- Added the gdtoa package from netlib's fp library so that ZDoom's printf-style
formatting can be entirely independant of the CRT.
SVN r1082 (trunk)
2008-07-23 04:57:26 +00:00
|
|
|
mysnprintf (buffer, countof(buffer), nametemplate, i + start);
|
2008-11-30 12:49:27 +00:00
|
|
|
|
|
|
|
lump = TexMan.CheckForTexture(buffer, FTexture::TEX_MiscPatch);
|
|
|
|
if (doomtemplate && lump.isValid() && i + start == 121)
|
2006-02-24 04:48:15 +00:00
|
|
|
{ // HACKHACK: Don't load STCFN121 in doom(2), because
|
2008-04-05 12:14:33 +00:00
|
|
|
// it's not really a lower-case 'y' but a '|'.
|
2006-02-24 04:48:15 +00:00
|
|
|
// Because a lot of wads with their own font seem to foolishly
|
2008-04-05 12:14:33 +00:00
|
|
|
// copy STCFN121 and make it a '|' themselves, wads must
|
2008-11-30 12:49:27 +00:00
|
|
|
// provide STCFN120 (x) and STCFN122 (z) for STCFN121 to load as a 'y'.
|
|
|
|
if (!TexMan.CheckForTexture("STCFN120", FTexture::TEX_MiscPatch).isValid() ||
|
|
|
|
!TexMan.CheckForTexture("STCFN122", FTexture::TEX_MiscPatch).isValid())
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-04-05 12:14:33 +00:00
|
|
|
// insert the incorrectly named '|' graphic in its correct position.
|
2008-11-30 12:49:27 +00:00
|
|
|
if (count > 124-start) charlumps[124-start] = TexMan[lump];
|
|
|
|
lump.SetInvalid();
|
2008-04-05 12:14:33 +00:00
|
|
|
stcfn121 = true;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
2008-04-05 12:14:33 +00:00
|
|
|
|
2008-11-30 12:49:27 +00:00
|
|
|
if (lump.isValid())
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
FTexture *pic = TexMan[lump];
|
2008-01-26 15:50:52 +00:00
|
|
|
if (pic != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-04-12 15:31:18 +00:00
|
|
|
// set the lump here only if it represents a valid texture
|
|
|
|
if (i != 124-start || !stcfn121)
|
2008-11-30 12:49:27 +00:00
|
|
|
charlumps[i] = pic;
|
2008-04-12 15:31:18 +00:00
|
|
|
|
2008-01-26 15:50:52 +00:00
|
|
|
int height = pic->GetScaledHeight();
|
|
|
|
int yoffs = pic->GetScaledTopOffset();
|
|
|
|
|
|
|
|
if (yoffs > maxyoffs)
|
|
|
|
{
|
|
|
|
maxyoffs = yoffs;
|
|
|
|
}
|
|
|
|
height += abs (yoffs);
|
|
|
|
if (height > FontHeight)
|
|
|
|
{
|
|
|
|
FontHeight = height;
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-11-30 12:49:27 +00:00
|
|
|
if (charlumps[i] != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
Chars[i].Pic = new FFontChar1 (charlumps[i]);
|
2010-02-26 05:34:30 +00:00
|
|
|
Chars[i].XMove = Chars[i].Pic->GetScaledWidth();
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
Chars[i].Pic = NULL;
|
2010-02-26 05:34:30 +00:00
|
|
|
Chars[i].XMove = INT_MIN;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-08-07 08:11:56 +00:00
|
|
|
if (spacewidth != -1)
|
|
|
|
{
|
|
|
|
SpaceWidth = spacewidth;
|
|
|
|
}
|
|
|
|
else if ('N'-first >= 0 && 'N'-first < count && Chars['N' - first].Pic != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
SpaceWidth = (Chars['N' - first].XMove + 1) / 2;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
SpaceWidth = 4;
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
|
|
|
|
FixXMoves();
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
LoadTranslations();
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
delete[] charlumps;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: ~FFont
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
FFont::~FFont ()
|
|
|
|
{
|
|
|
|
if (Chars)
|
|
|
|
{
|
|
|
|
int count = LastChar - FirstChar + 1;
|
|
|
|
|
|
|
|
for (int i = 0; i < count; ++i)
|
|
|
|
{
|
2006-05-09 03:40:15 +00:00
|
|
|
if (Chars[i].Pic != NULL && Chars[i].Pic->Name[0] == 0)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
delete Chars[i].Pic;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
delete[] Chars;
|
|
|
|
Chars = NULL;
|
|
|
|
}
|
|
|
|
if (PatchRemap)
|
|
|
|
{
|
|
|
|
delete[] PatchRemap;
|
|
|
|
PatchRemap = NULL;
|
|
|
|
}
|
2006-05-09 03:40:15 +00:00
|
|
|
if (Name)
|
|
|
|
{
|
|
|
|
delete[] Name;
|
|
|
|
Name = NULL;
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
FFont **prev = &FirstFont;
|
|
|
|
FFont *font = *prev;
|
|
|
|
|
|
|
|
while (font != NULL && font != this)
|
|
|
|
{
|
|
|
|
prev = &font->Next;
|
|
|
|
font = *prev;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (font != NULL)
|
|
|
|
{
|
|
|
|
*prev = font->Next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: FindFont
|
|
|
|
//
|
|
|
|
// Searches for the named font in the list of loaded fonts, returning the
|
|
|
|
// font if it was found. The disk is not checked if it cannot be found.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
FFont *FFont::FindFont (const char *name)
|
|
|
|
{
|
|
|
|
if (name == NULL)
|
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
FFont *font = FirstFont;
|
|
|
|
|
|
|
|
while (font != NULL)
|
|
|
|
{
|
|
|
|
if (stricmp (font->Name, name) == 0)
|
|
|
|
break;
|
|
|
|
font = font->Next;
|
|
|
|
}
|
|
|
|
return font;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// RecordTextureColors
|
|
|
|
//
|
|
|
|
// Given a 256 entry buffer, sets every entry that corresponds to a color
|
|
|
|
// used by the texture to 1.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-09-14 00:02:31 +00:00
|
|
|
void RecordTextureColors (FTexture *pic, BYTE *usedcolors)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
int x;
|
|
|
|
|
|
|
|
for (x = pic->GetWidth() - 1; x >= 0; x--)
|
|
|
|
{
|
|
|
|
const FTexture::Span *spans;
|
|
|
|
const BYTE *column = pic->GetColumn (x, &spans);
|
|
|
|
|
|
|
|
while (spans->Length != 0)
|
|
|
|
{
|
|
|
|
const BYTE *source = column + spans->TopOffset;
|
|
|
|
int count = spans->Length;
|
|
|
|
|
|
|
|
do
|
|
|
|
{
|
|
|
|
usedcolors[*source++] = 1;
|
|
|
|
} while (--count);
|
|
|
|
|
|
|
|
spans++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// compare
|
|
|
|
//
|
|
|
|
// Used for sorting colors by brightness.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
static int STACK_ARGS compare (const void *arg1, const void *arg2)
|
|
|
|
{
|
2006-09-14 00:02:31 +00:00
|
|
|
if (RPART(GPalette.BaseColors[*((BYTE *)arg1)]) * 299 +
|
|
|
|
GPART(GPalette.BaseColors[*((BYTE *)arg1)]) * 587 +
|
|
|
|
BPART(GPalette.BaseColors[*((BYTE *)arg1)]) * 114 <
|
|
|
|
RPART(GPalette.BaseColors[*((BYTE *)arg2)]) * 299 +
|
|
|
|
GPART(GPalette.BaseColors[*((BYTE *)arg2)]) * 587 +
|
|
|
|
BPART(GPalette.BaseColors[*((BYTE *)arg2)]) * 114)
|
2006-02-24 04:48:15 +00:00
|
|
|
return -1;
|
|
|
|
else
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: SimpleTranslation
|
|
|
|
//
|
|
|
|
// Colorsused, translation, and reverse must all be 256 entry buffers.
|
|
|
|
// Colorsused must already be filled out.
|
|
|
|
// Translation be set to remap the source colors to a new range of
|
|
|
|
// consecutive colors based at 1 (0 is transparent).
|
|
|
|
// Reverse will be just the opposite of translation: It maps the new color
|
|
|
|
// range to the original colors.
|
|
|
|
// *Luminosity will be an array just large enough to hold the brightness
|
|
|
|
// levels of all the used colors, in consecutive order. It is sorted from
|
|
|
|
// darkest to lightest and scaled such that the darkest color is 0.0 and
|
|
|
|
// the brightest color is 1.0.
|
|
|
|
// The return value is the number of used colors and thus the number of
|
|
|
|
// entries in *luminosity.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-09-14 00:02:31 +00:00
|
|
|
int FFont::SimpleTranslation (BYTE *colorsused, BYTE *translation, BYTE *reverse, double **luminosity)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
double min, max, diver;
|
|
|
|
int i, j;
|
|
|
|
|
|
|
|
memset (translation, 0, 256);
|
|
|
|
|
|
|
|
reverse[0] = 0;
|
2006-04-11 16:27:41 +00:00
|
|
|
for (i = 1, j = 1; i < 256; i++)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (colorsused[i])
|
|
|
|
{
|
|
|
|
reverse[j++] = i;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
qsort (reverse+1, j-1, 1, compare);
|
|
|
|
|
|
|
|
*luminosity = new double[j];
|
2013-05-26 04:03:47 +00:00
|
|
|
(*luminosity)[0] = 0.0; // [BL] Prevent uninitalized memory
|
2006-02-24 04:48:15 +00:00
|
|
|
max = 0.0;
|
|
|
|
min = 100000000.0;
|
|
|
|
for (i = 1; i < j; i++)
|
|
|
|
{
|
|
|
|
translation[reverse[i]] = i;
|
|
|
|
|
|
|
|
(*luminosity)[i] = RPART(GPalette.BaseColors[reverse[i]]) * 0.299 +
|
|
|
|
GPART(GPalette.BaseColors[reverse[i]]) * 0.587 +
|
|
|
|
BPART(GPalette.BaseColors[reverse[i]]) * 0.114;
|
|
|
|
if ((*luminosity)[i] > max)
|
|
|
|
max = (*luminosity)[i];
|
|
|
|
if ((*luminosity)[i] < min)
|
|
|
|
min = (*luminosity)[i];
|
|
|
|
}
|
|
|
|
diver = 1.0 / (max - min);
|
2006-04-11 16:27:41 +00:00
|
|
|
for (i = 1; i < j; i++)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
(*luminosity)[i] = ((*luminosity)[i] - min) * diver;
|
|
|
|
}
|
|
|
|
|
|
|
|
return j;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: BuildTranslations
|
|
|
|
//
|
|
|
|
// Build color translations for this font. Luminosity is an array of
|
|
|
|
// brightness levels. The ActiveColors member must be set to indicate how
|
|
|
|
// large this array is. Identity is an array that remaps the colors to
|
|
|
|
// their original values; it is only used for CR_UNTRANSLATED. Ranges
|
|
|
|
// is an array of TranslationParm structs defining the ranges for every
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
// possible color, in order. Palette is the colors to use for the
|
|
|
|
// untranslated version of the font.
|
2006-08-30 02:38:39 +00:00
|
|
|
//
|
|
|
|
//==========================================================================
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
void FFont::BuildTranslations (const double *luminosity, const BYTE *identity,
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
const void *ranges, int total_colors, const PalEntry *palette)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2006-08-30 02:38:39 +00:00
|
|
|
int i, j;
|
|
|
|
const TranslationParm *parmstart = (const TranslationParm *)ranges;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
FRemapTable remap(total_colors);
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
// Create different translations for different color ranges
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
Ranges.Clear();
|
2006-08-30 02:38:39 +00:00
|
|
|
for (i = 0; i < NumTextColors; i++)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
if (i == CR_UNTRANSLATED)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2006-08-30 02:38:39 +00:00
|
|
|
if (identity != NULL)
|
|
|
|
{
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
memcpy (remap.Remap, identity, ActiveColors);
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (palette != NULL)
|
|
|
|
{
|
|
|
|
memcpy (remap.Palette, palette, ActiveColors*sizeof(PalEntry));
|
|
|
|
}
|
|
|
|
else
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
remap.Palette[0] = GPalette.BaseColors[identity[0]] & MAKEARGB(0,255,255,255);
|
|
|
|
for (j = 1; j < ActiveColors; ++j)
|
|
|
|
{
|
|
|
|
remap.Palette[j] = GPalette.BaseColors[identity[j]] | MAKEARGB(255,0,0,0);
|
|
|
|
}
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
remap = Ranges[0];
|
2007-12-23 16:53:39 +00:00
|
|
|
}
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
Ranges.Push(remap);
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
continue;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
assert(parmstart->RangeStart >= 0);
|
|
|
|
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
remap.Remap[0] = 0;
|
|
|
|
remap.Palette[0] = 0;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
for (j = 1; j < ActiveColors; j++)
|
|
|
|
{
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
int v = int(luminosity[j] * 256.0);
|
2006-02-24 04:48:15 +00:00
|
|
|
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
// Find the color range that this luminosity value lies within.
|
2006-08-30 02:38:39 +00:00
|
|
|
const TranslationParm *parms = parmstart - 1;
|
|
|
|
do
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2006-08-30 02:38:39 +00:00
|
|
|
parms++;
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
if (parms->RangeStart <= v && parms->RangeEnd >= v)
|
|
|
|
break;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
while (parms[1].RangeStart > parms[0].RangeEnd);
|
|
|
|
|
- Added support for casting color names in ACS print-type statements. What that
means is that instead of writing this:
print (s:"\cDSome text");
You can write this:
print (r:CR_GREEN, s:"Some text");
- Added some new colors, based on the ones jimmy91 posted in the forum: cream,
light blue, black, olive, dark green, dark red, dark brown, purple, and dark
gray.
- Simplified FFont::BuildTranslations() and BuildTranslations2() to make adding
new colors easier.
SVN r306 (trunk)
2006-08-22 21:58:24 +00:00
|
|
|
// Linearly interpolate to find out which color this luminosity level gets.
|
|
|
|
int rangev = ((v - parms->RangeStart) << 8) / (parms->RangeEnd - parms->RangeStart);
|
|
|
|
int r = ((parms->Start[0] << 8) + rangev * (parms->End[0] - parms->Start[0])) >> 8; // red
|
|
|
|
int g = ((parms->Start[1] << 8) + rangev * (parms->End[1] - parms->Start[1])) >> 8; // green
|
|
|
|
int b = ((parms->Start[2] << 8) + rangev * (parms->End[2] - parms->Start[2])) >> 8; // blue
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
r = clamp(r, 0, 255);
|
|
|
|
g = clamp(g, 0, 255);
|
|
|
|
b = clamp(b, 0, 255);
|
|
|
|
remap.Remap[j] = ColorMatcher.Pick(r, g, b);
|
|
|
|
remap.Palette[j] = PalEntry(255,r,g,b);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
Ranges.Push(remap);
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
// Advance to the next color range.
|
|
|
|
while (parmstart[1].RangeStart > parmstart[0].RangeEnd)
|
|
|
|
{
|
|
|
|
parmstart++;
|
|
|
|
}
|
|
|
|
parmstart++;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: GetColorTranslation
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
FRemapTable *FFont::GetColorTranslation (EColorRange range) const
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (ActiveColors == 0)
|
|
|
|
return NULL;
|
2006-08-30 02:38:39 +00:00
|
|
|
else if (range >= NumTextColors)
|
|
|
|
range = CR_UNTRANSLATED;
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
return &Ranges[range];
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
2010-02-26 05:34:30 +00:00
|
|
|
// FFont :: GetCharCode
|
|
|
|
//
|
|
|
|
// If the character code is in the font, returns it. If it is not, but it
|
|
|
|
// is lowercase and has an uppercase variant present, return that. Otherwise
|
|
|
|
// return -1.
|
2006-08-30 02:38:39 +00:00
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
int FFont::GetCharCode(int code, bool needpic) const
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-04-12 17:33:55 +00:00
|
|
|
if (code < 0 && code >= -128)
|
|
|
|
{
|
|
|
|
// regular chars turn negative when the 8th bit is set.
|
|
|
|
code &= 255;
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
if (code >= FirstChar && code <= LastChar && (!needpic || Chars[code - FirstChar].Pic != NULL))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
return code;
|
|
|
|
}
|
2013-05-17 02:35:35 +00:00
|
|
|
// Try converting lowercase characters to uppercase.
|
|
|
|
if (myislower(code))
|
2010-02-26 05:34:30 +00:00
|
|
|
{
|
|
|
|
code -= 32;
|
|
|
|
if (code >= FirstChar && code <= LastChar && (!needpic || Chars[code - FirstChar].Pic != NULL))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
return code;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
2013-05-17 02:35:35 +00:00
|
|
|
// Try stripping accents from accented characters.
|
|
|
|
int newcode = stripaccent(code);
|
|
|
|
if (newcode != code)
|
|
|
|
{
|
|
|
|
code = newcode;
|
|
|
|
if (code >= FirstChar && code <= LastChar && (!needpic || Chars[code - FirstChar].Pic != NULL))
|
|
|
|
{
|
|
|
|
return code;
|
|
|
|
}
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
return -1;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
2010-02-26 05:34:30 +00:00
|
|
|
// FFont :: GetChar
|
2006-08-30 02:38:39 +00:00
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
FTexture *FFont::GetChar (int code, int *const width) const
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
code = GetCharCode(code, false);
|
|
|
|
int xmove = SpaceWidth;
|
|
|
|
|
|
|
|
if (code >= 0)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
code -= FirstChar;
|
|
|
|
xmove = Chars[code].XMove;
|
|
|
|
if (Chars[code].Pic == NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
code = GetCharCode(code + FirstChar, true);
|
|
|
|
if (code >= 0)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
code -= FirstChar;
|
|
|
|
xmove = Chars[code].XMove;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
if (width != NULL)
|
|
|
|
{
|
|
|
|
*width = xmove;
|
|
|
|
}
|
|
|
|
return (code < 0) ? NULL : Chars[code].Pic;
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: GetCharWidth
|
|
|
|
//
|
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
int FFont::GetCharWidth (int code) const
|
|
|
|
{
|
|
|
|
code = GetCharCode(code, false);
|
|
|
|
return (code < 0) ? SpaceWidth : Chars[code - FirstChar].XMove;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: LoadTranslations
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFont::LoadTranslations()
|
|
|
|
{
|
|
|
|
unsigned int count = LastChar - FirstChar + 1;
|
|
|
|
BYTE usedcolors[256], identity[256];
|
|
|
|
double *luminosity;
|
|
|
|
|
|
|
|
memset (usedcolors, 0, 256);
|
|
|
|
for (unsigned int i = 0; i < count; i++)
|
|
|
|
{
|
|
|
|
FFontChar1 *pic = static_cast<FFontChar1 *>(Chars[i].Pic);
|
|
|
|
if(pic)
|
|
|
|
{
|
|
|
|
pic->SetSourceRemap(NULL); // Force the FFontChar1 to return the same pixels as the base texture
|
|
|
|
RecordTextureColors (pic, usedcolors);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
ActiveColors = SimpleTranslation (usedcolors, PatchRemap, identity, &luminosity);
|
|
|
|
|
|
|
|
for (unsigned int i = 0; i < count; i++)
|
|
|
|
{
|
|
|
|
if(Chars[i].Pic)
|
|
|
|
static_cast<FFontChar1 *>(Chars[i].Pic)->SetSourceRemap(PatchRemap);
|
|
|
|
}
|
|
|
|
|
|
|
|
BuildTranslations (luminosity, identity, &TranslationParms[0][0], ActiveColors, NULL);
|
|
|
|
|
|
|
|
delete[] luminosity;
|
|
|
|
}
|
|
|
|
|
2008-01-27 05:23:10 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: Preload
|
|
|
|
//
|
|
|
|
// Loads most of the 7-bit ASCII characters. In the case of D3DFB, this
|
|
|
|
// means all the characters of a font have a better chance of being packed
|
|
|
|
// into the same hardware texture.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFont::Preload() const
|
|
|
|
{
|
|
|
|
// First and last char are the same? Wait until it's actually needed
|
|
|
|
// since nothing is gained by preloading now.
|
|
|
|
if (FirstChar == LastChar)
|
|
|
|
{
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
for (int i = MAX(FirstChar, 0x21); i < MIN(LastChar, 0x7e); ++i)
|
|
|
|
{
|
|
|
|
int foo;
|
|
|
|
FTexture *pic = GetChar(i, &foo);
|
|
|
|
if (pic != NULL)
|
|
|
|
{
|
|
|
|
pic->GetNative(false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: StaticPreloadFonts
|
|
|
|
//
|
|
|
|
// Preloads all the defined fonts.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFont::StaticPreloadFonts()
|
|
|
|
{
|
|
|
|
for (FFont *font = FirstFont; font != NULL; font = font->Next)
|
|
|
|
{
|
|
|
|
font->Preload();
|
|
|
|
}
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: FFont - default constructor
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FFont::FFont (int lump)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
Lump = lump;
|
2006-02-24 04:48:15 +00:00
|
|
|
Chars = NULL;
|
|
|
|
PatchRemap = NULL;
|
2006-05-09 03:40:15 +00:00
|
|
|
Name = NULL;
|
2013-04-30 03:14:49 +00:00
|
|
|
Cursor = '_';
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: FSingleLumpFont
|
|
|
|
//
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
// Loads a FON1 or FON2 font resource.
|
2006-08-30 02:38:39 +00:00
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FSingleLumpFont::FSingleLumpFont (const char *name, int lump) : FFont(lump)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
assert(lump >= 0);
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
Name = copystring (name);
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
FMemLump data1 = Wads.ReadLump (lump);
|
|
|
|
const BYTE *data = (const BYTE *)data1.GetMem();
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
if (data[0] == 0xE1 && data[1] == 0xE6 && data[2] == 0xD5 && data[3] == 0x1A)
|
|
|
|
{
|
|
|
|
LoadBMF(lump, data);
|
|
|
|
}
|
|
|
|
else if (data[0] != 'F' || data[1] != 'O' || data[2] != 'N' ||
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
(data[3] != '1' && data[3] != '2'))
|
|
|
|
{
|
|
|
|
I_FatalError ("%s is not a recognizable font", name);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
switch (data[3])
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
case '1':
|
|
|
|
LoadFON1 (lump, data);
|
|
|
|
break;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
case '2':
|
|
|
|
LoadFON2 (lump, data);
|
|
|
|
break;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
Next = FirstFont;
|
|
|
|
FirstFont = this;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: CreateFontFromPic
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2008-06-15 18:36:26 +00:00
|
|
|
void FSingleLumpFont::CreateFontFromPic (FTextureID picnum)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
FTexture *pic = TexMan[picnum];
|
|
|
|
|
|
|
|
FontHeight = pic->GetHeight ();
|
|
|
|
SpaceWidth = pic->GetWidth ();
|
|
|
|
GlobalKerning = 0;
|
|
|
|
|
|
|
|
FirstChar = LastChar = 'A';
|
|
|
|
Chars = new CharData[1];
|
|
|
|
Chars->Pic = pic;
|
|
|
|
|
|
|
|
// Only one color range. Don't bother with the others.
|
|
|
|
ActiveColors = 0;
|
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: LoadTranslations
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FSingleLumpFont::LoadTranslations()
|
|
|
|
{
|
|
|
|
double luminosity[256];
|
|
|
|
BYTE identity[256];
|
|
|
|
PalEntry local_palette[256];
|
|
|
|
bool useidentity = true;
|
|
|
|
bool usepalette = false;
|
|
|
|
const void* ranges;
|
|
|
|
unsigned int count = LastChar - FirstChar + 1;
|
|
|
|
|
|
|
|
switch(FontType)
|
|
|
|
{
|
|
|
|
case FONT1:
|
|
|
|
useidentity = false;
|
|
|
|
ranges = &TranslationParms[1][0];
|
|
|
|
CheckFON1Chars (luminosity);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case BMFFONT:
|
|
|
|
case FONT2:
|
|
|
|
usepalette = true;
|
|
|
|
FixupPalette (identity, luminosity, PaletteData, RescalePalette, local_palette);
|
|
|
|
|
|
|
|
ranges = &TranslationParms[0][0];
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
2016-01-24 00:33:26 +00:00
|
|
|
// Should be unreachable.
|
|
|
|
I_Error("Unknown font type in FSingleLumpFont::LoadTranslation.");
|
|
|
|
return;
|
2012-08-06 10:25:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
for(unsigned int i = 0;i < count;++i)
|
|
|
|
{
|
|
|
|
if(Chars[i].Pic)
|
|
|
|
static_cast<FFontChar2*>(Chars[i].Pic)->SetSourceRemap(PatchRemap);
|
|
|
|
}
|
|
|
|
|
|
|
|
BuildTranslations (luminosity, useidentity ? identity : NULL, ranges, ActiveColors, usepalette ? local_palette : NULL);
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: LoadFON1
|
|
|
|
//
|
|
|
|
// FON1 is used for the console font.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FSingleLumpFont::LoadFON1 (int lump, const BYTE *data)
|
|
|
|
{
|
2006-08-30 02:38:39 +00:00
|
|
|
int w, h;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
Chars = new CharData[256];
|
|
|
|
|
|
|
|
w = data[4] + data[5]*256;
|
|
|
|
h = data[6] + data[7]*256;
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FontType = FONT1;
|
2006-02-24 04:48:15 +00:00
|
|
|
FontHeight = h;
|
|
|
|
SpaceWidth = w;
|
|
|
|
FirstChar = 0;
|
|
|
|
LastChar = 255;
|
|
|
|
GlobalKerning = 0;
|
2006-08-30 02:38:39 +00:00
|
|
|
PatchRemap = new BYTE[256];
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
for(unsigned int i = 0;i < 256;++i)
|
|
|
|
Chars[i].Pic = NULL;
|
|
|
|
|
|
|
|
LoadTranslations();
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: LoadFON2
|
|
|
|
//
|
|
|
|
// FON2 is used for everything but the console font. The console font should
|
|
|
|
// probably use FON2 as well, but oh well.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FSingleLumpFont::LoadFON2 (int lump, const BYTE *data)
|
|
|
|
{
|
|
|
|
int count, i, totalwidth;
|
|
|
|
int *widths2;
|
|
|
|
WORD *widths;
|
|
|
|
const BYTE *palette;
|
|
|
|
const BYTE *data_p;
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FontType = FONT2;
|
2006-02-24 04:48:15 +00:00
|
|
|
FontHeight = data[4] + data[5]*256;
|
|
|
|
FirstChar = data[6];
|
|
|
|
LastChar = data[7];
|
2012-11-30 23:36:02 +00:00
|
|
|
ActiveColors = data[10]+1;
|
2012-08-06 10:25:51 +00:00
|
|
|
PatchRemap = NULL;
|
|
|
|
RescalePalette = data[9] == 0;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
count = LastChar - FirstChar + 1;
|
|
|
|
Chars = new CharData[count];
|
|
|
|
widths2 = new int[count];
|
|
|
|
if (data[11] & 1)
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{ // Font specifies a kerning value.
|
2006-02-24 04:48:15 +00:00
|
|
|
GlobalKerning = LittleShort(*(SWORD *)&data[12]);
|
|
|
|
widths = (WORD *)(data + 14);
|
|
|
|
}
|
|
|
|
else
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{ // Font does not specify a kerning value.
|
2006-02-24 04:48:15 +00:00
|
|
|
GlobalKerning = 0;
|
|
|
|
widths = (WORD *)(data + 12);
|
|
|
|
}
|
|
|
|
totalwidth = 0;
|
|
|
|
|
|
|
|
if (data[8])
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{ // Font is mono-spaced.
|
2006-02-24 04:48:15 +00:00
|
|
|
totalwidth = LittleShort(widths[0]);
|
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
{
|
|
|
|
widths2[i] = totalwidth;
|
|
|
|
}
|
|
|
|
totalwidth *= count;
|
|
|
|
palette = (BYTE *)&widths[1];
|
|
|
|
}
|
|
|
|
else
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{ // Font has varying character widths.
|
2006-02-24 04:48:15 +00:00
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
{
|
|
|
|
widths2[i] = LittleShort(widths[i]);
|
|
|
|
totalwidth += widths2[i];
|
|
|
|
}
|
|
|
|
palette = (BYTE *)(widths + i);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (FirstChar <= ' ' && LastChar >= ' ')
|
|
|
|
{
|
|
|
|
SpaceWidth = widths2[' '-FirstChar];
|
|
|
|
}
|
|
|
|
else if (FirstChar <= 'N' && LastChar >= 'N')
|
|
|
|
{
|
|
|
|
SpaceWidth = (widths2['N' - FirstChar] + 1) / 2;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
SpaceWidth = totalwidth * 2 / (3 * count);
|
|
|
|
}
|
|
|
|
|
2012-11-30 23:36:02 +00:00
|
|
|
memcpy(PaletteData, palette, ActiveColors*3);
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-11-30 23:36:02 +00:00
|
|
|
data_p = palette + ActiveColors*3;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
{
|
|
|
|
int destSize = widths2[i] * FontHeight;
|
2010-02-26 05:34:30 +00:00
|
|
|
Chars[i].XMove = widths2[i];
|
2006-02-24 04:48:15 +00:00
|
|
|
if (destSize <= 0)
|
|
|
|
{
|
|
|
|
Chars[i].Pic = NULL;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
Chars[i].Pic = new FFontChar2 (lump, int(data_p - data), widths2[i], FontHeight);
|
2006-02-24 04:48:15 +00:00
|
|
|
do
|
|
|
|
{
|
|
|
|
SBYTE code = *data_p++;
|
|
|
|
if (code >= 0)
|
|
|
|
{
|
|
|
|
data_p += code+1;
|
|
|
|
destSize -= code+1;
|
|
|
|
}
|
|
|
|
else if (code != -128)
|
|
|
|
{
|
|
|
|
data_p++;
|
|
|
|
destSize -= (-code)+1;
|
|
|
|
}
|
|
|
|
} while (destSize > 0);
|
|
|
|
}
|
|
|
|
if (destSize < 0)
|
|
|
|
{
|
|
|
|
i += FirstChar;
|
|
|
|
I_FatalError ("Overflow decompressing char %d (%c) of %s", i, i, Name);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
LoadTranslations();
|
2006-05-19 05:14:37 +00:00
|
|
|
delete[] widths2;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: LoadBMF
|
|
|
|
//
|
|
|
|
// Loads a BMF font. The file format is described at
|
|
|
|
// <http://bmf.wz.cz/bmf-format.htm>
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FSingleLumpFont::LoadBMF(int lump, const BYTE *data)
|
|
|
|
{
|
|
|
|
const BYTE *chardata;
|
|
|
|
int numchars, count, totalwidth, nwidth;
|
|
|
|
int infolen;
|
|
|
|
int i, chari;
|
|
|
|
BYTE raw_palette[256*3];
|
|
|
|
PalEntry sort_palette[256];
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FontType = BMFFONT;
|
2010-02-26 05:34:30 +00:00
|
|
|
FontHeight = data[5];
|
|
|
|
GlobalKerning = (SBYTE)data[8];
|
|
|
|
ActiveColors = data[16];
|
|
|
|
SpaceWidth = -1;
|
|
|
|
nwidth = -1;
|
2012-08-06 10:25:51 +00:00
|
|
|
RescalePalette = true;
|
2010-02-26 05:34:30 +00:00
|
|
|
|
|
|
|
infolen = data[17 + ActiveColors*3];
|
|
|
|
chardata = data + 18 + ActiveColors*3 + infolen;
|
|
|
|
numchars = chardata[0] + 256*chardata[1];
|
|
|
|
chardata += 2;
|
|
|
|
|
|
|
|
// Scan for lowest and highest characters defined and total font width.
|
|
|
|
FirstChar = 256;
|
|
|
|
LastChar = 0;
|
|
|
|
totalwidth = 0;
|
|
|
|
for (i = chari = 0; i < numchars; ++i, chari += 6 + chardata[chari+1] * chardata[chari+2])
|
|
|
|
{
|
|
|
|
if ((chardata[chari+1] == 0 || chardata[chari+2] == 0) && chardata[chari+5] == 0)
|
|
|
|
{ // Don't count empty characters.
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (chardata[chari] < FirstChar)
|
|
|
|
{
|
|
|
|
FirstChar = chardata[chari];
|
|
|
|
}
|
|
|
|
if (chardata[chari] > LastChar)
|
|
|
|
{
|
|
|
|
LastChar = chardata[chari];
|
|
|
|
}
|
|
|
|
totalwidth += chardata[chari+1];
|
|
|
|
}
|
|
|
|
if (LastChar < FirstChar)
|
|
|
|
{
|
|
|
|
I_FatalError("BMF font defines no characters");
|
|
|
|
}
|
|
|
|
count = LastChar - FirstChar + 1;
|
|
|
|
Chars = new CharData[count];
|
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
{
|
|
|
|
Chars[i].Pic = NULL;
|
|
|
|
Chars[i].XMove = INT_MIN;
|
|
|
|
}
|
|
|
|
|
|
|
|
// BMF palettes are only six bits per component. Fix that.
|
|
|
|
for (i = 0; i < ActiveColors*3; ++i)
|
|
|
|
{
|
2011-06-17 22:46:34 +00:00
|
|
|
raw_palette[i+3] = (data[17 + i] << 2) | (data[17 + i] >> 4);
|
2010-02-26 05:34:30 +00:00
|
|
|
}
|
2011-06-17 22:46:34 +00:00
|
|
|
ActiveColors++;
|
2010-02-26 05:34:30 +00:00
|
|
|
|
|
|
|
// Sort the palette by increasing brightness
|
|
|
|
for (i = 0; i < ActiveColors; ++i)
|
|
|
|
{
|
|
|
|
PalEntry *pal = &sort_palette[i];
|
|
|
|
pal->a = i; // Use alpha part to point back to original entry
|
|
|
|
pal->r = raw_palette[i*3 + 0];
|
|
|
|
pal->g = raw_palette[i*3 + 1];
|
|
|
|
pal->b = raw_palette[i*3 + 2];
|
|
|
|
}
|
|
|
|
qsort(sort_palette + 1, ActiveColors - 1, sizeof(PalEntry), BMFCompare);
|
|
|
|
|
|
|
|
// Create the PatchRemap table from the sorted "alpha" values.
|
|
|
|
PatchRemap = new BYTE[ActiveColors];
|
|
|
|
PatchRemap[0] = 0;
|
|
|
|
for (i = 1; i < ActiveColors; ++i)
|
|
|
|
{
|
|
|
|
PatchRemap[sort_palette[i].a] = i;
|
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
memcpy(PaletteData, raw_palette, 768);
|
2010-02-26 05:34:30 +00:00
|
|
|
|
|
|
|
// Now scan through the characters again, creating glyphs for each one.
|
|
|
|
for (i = chari = 0; i < numchars; ++i, chari += 6 + chardata[chari+1] * chardata[chari+2])
|
|
|
|
{
|
|
|
|
assert(chardata[chari] - FirstChar >= 0);
|
|
|
|
assert(chardata[chari] - FirstChar < count);
|
|
|
|
if (chardata[chari] == ' ')
|
|
|
|
{
|
|
|
|
SpaceWidth = chardata[chari+5];
|
|
|
|
}
|
|
|
|
else if (chardata[chari] == 'N')
|
|
|
|
{
|
|
|
|
nwidth = chardata[chari+5];
|
|
|
|
}
|
|
|
|
Chars[chardata[chari] - FirstChar].XMove = chardata[chari+5];
|
|
|
|
if (chardata[chari+1] == 0 || chardata[chari+2] == 0)
|
|
|
|
{ // Empty character: skip it.
|
|
|
|
continue;
|
|
|
|
}
|
2012-08-06 10:25:51 +00:00
|
|
|
Chars[chardata[chari] - FirstChar].Pic = new FFontChar2(lump, int(chardata + chari + 6 - data),
|
2010-02-26 05:34:30 +00:00
|
|
|
chardata[chari+1], // width
|
|
|
|
chardata[chari+2], // height
|
|
|
|
-(SBYTE)chardata[chari+3], // x offset
|
|
|
|
-(SBYTE)chardata[chari+4] // y offset
|
|
|
|
);
|
|
|
|
}
|
|
|
|
|
|
|
|
// If the font did not define a space character, determine a suitable space width now.
|
|
|
|
if (SpaceWidth < 0)
|
|
|
|
{
|
|
|
|
if (nwidth >= 0)
|
|
|
|
{
|
|
|
|
SpaceWidth = nwidth;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
SpaceWidth = totalwidth * 2 / (3 * count);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
FixXMoves();
|
2012-08-06 10:25:51 +00:00
|
|
|
LoadTranslations();
|
2010-02-26 05:34:30 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: BMFCompare STATIC
|
|
|
|
//
|
|
|
|
// Helper to sort BMF palettes.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
int STACK_ARGS FSingleLumpFont::BMFCompare(const void *a, const void *b)
|
|
|
|
{
|
|
|
|
const PalEntry *pa = (const PalEntry *)a;
|
|
|
|
const PalEntry *pb = (const PalEntry *)b;
|
|
|
|
|
|
|
|
return (pa->r * 299 + pa->g * 587 + pa->b * 114) -
|
|
|
|
(pb->r * 299 + pb->g * 587 + pb->b * 114);
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: CheckFON1Chars
|
|
|
|
//
|
|
|
|
// Scans a FON1 resource for all the color values it uses and sets up
|
|
|
|
// some tables like SimpleTranslation. Data points to the RLE data for
|
|
|
|
// the characters. Also sets up the character textures.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
void FSingleLumpFont::CheckFON1Chars (double *luminosity)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
FMemLump memLump = Wads.ReadLump(Lump);
|
|
|
|
const BYTE* data = (const BYTE*) memLump.GetMem();
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
BYTE used[256], reverse[256];
|
|
|
|
const BYTE *data_p;
|
|
|
|
int i, j;
|
|
|
|
|
|
|
|
memset (used, 0, 256);
|
|
|
|
data_p = data + 8;
|
|
|
|
|
|
|
|
for (i = 0; i < 256; ++i)
|
|
|
|
{
|
|
|
|
int destSize = SpaceWidth * FontHeight;
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
if(!Chars[i].Pic)
|
|
|
|
{
|
|
|
|
Chars[i].Pic = new FFontChar2 (Lump, int(data_p - data), SpaceWidth, FontHeight);
|
|
|
|
Chars[i].XMove = SpaceWidth;
|
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
// Advance to next char's data and count the used colors.
|
|
|
|
do
|
|
|
|
{
|
|
|
|
SBYTE code = *data_p++;
|
|
|
|
if (code >= 0)
|
|
|
|
{
|
|
|
|
destSize -= code+1;
|
|
|
|
while (code-- >= 0)
|
|
|
|
{
|
|
|
|
used[*data_p++] = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (code != -128)
|
|
|
|
{
|
|
|
|
used[*data_p++] = 1;
|
|
|
|
destSize -= 1 - code;
|
|
|
|
}
|
|
|
|
} while (destSize > 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
memset (PatchRemap, 0, 256);
|
|
|
|
reverse[0] = 0;
|
|
|
|
for (i = 1, j = 1; i < 256; ++i)
|
|
|
|
{
|
|
|
|
if (used[i])
|
|
|
|
{
|
|
|
|
reverse[j++] = i;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (i = 1; i < j; ++i)
|
|
|
|
{
|
|
|
|
PatchRemap[reverse[i]] = i;
|
|
|
|
luminosity[i] = (reverse[i] - 1) / 254.0;
|
|
|
|
}
|
|
|
|
ActiveColors = j;
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSingleLumpFont :: FixupPalette
|
|
|
|
//
|
|
|
|
// Finds the best matches for the colors used by a FON2 font and sets up
|
|
|
|
// some tables like SimpleTranslation.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
void FSingleLumpFont::FixupPalette (BYTE *identity, double *luminosity, const BYTE *palette, bool rescale, PalEntry *out_palette)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
double maxlum = 0.0;
|
|
|
|
double minlum = 100000000.0;
|
|
|
|
double diver;
|
|
|
|
|
|
|
|
identity[0] = 0;
|
|
|
|
palette += 3; // Skip the transparent color
|
|
|
|
|
2012-08-06 09:49:15 +00:00
|
|
|
for (i = 1; i < ActiveColors; ++i, palette += 3)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
int r = palette[0];
|
|
|
|
int g = palette[1];
|
|
|
|
int b = palette[2];
|
|
|
|
double lum = r*0.299 + g*0.587 + b*0.114;
|
|
|
|
identity[i] = ColorMatcher.Pick (r, g, b);
|
|
|
|
luminosity[i] = lum;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
out_palette[i].r = r;
|
|
|
|
out_palette[i].g = g;
|
|
|
|
out_palette[i].b = b;
|
|
|
|
out_palette[i].a = 255;
|
2006-02-24 04:48:15 +00:00
|
|
|
if (lum > maxlum)
|
|
|
|
maxlum = lum;
|
|
|
|
if (lum < minlum)
|
|
|
|
minlum = lum;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
out_palette[0] = 0;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
if (rescale)
|
|
|
|
{
|
|
|
|
diver = 1.0 / (maxlum - minlum);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
diver = 1.0 / 255.0;
|
|
|
|
}
|
2012-08-06 09:49:15 +00:00
|
|
|
for (i = 1; i < ActiveColors; ++i)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
luminosity[i] = (luminosity[i] - minlum) * diver;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSinglePicFont :: FSinglePicFont
|
|
|
|
//
|
|
|
|
// Creates a font to wrap a texture so that you can use hudmessage as if it
|
|
|
|
// were a hudpic command. It does not support translation, but animation
|
|
|
|
// is supported, unlike all the real fonts.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FSinglePicFont::FSinglePicFont(const char *picname) :
|
|
|
|
FFont(-1) // Since lump is only needed for priority information we don't need to worry about this here.
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{
|
2008-06-15 18:36:26 +00:00
|
|
|
FTextureID picnum = TexMan.CheckForTexture (picname, FTexture::TEX_Any);
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
|
2008-06-15 18:36:26 +00:00
|
|
|
if (!picnum.isValid())
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
{
|
|
|
|
I_FatalError ("%s is not a font or texture", picname);
|
|
|
|
}
|
|
|
|
|
|
|
|
FTexture *pic = TexMan[picnum];
|
|
|
|
|
|
|
|
Name = copystring(picname);
|
2010-03-05 03:31:20 +00:00
|
|
|
FontHeight = pic->GetScaledHeight();
|
|
|
|
SpaceWidth = pic->GetScaledWidth();
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
GlobalKerning = 0;
|
|
|
|
FirstChar = LastChar = 'A';
|
|
|
|
ActiveColors = 0;
|
|
|
|
PicNum = picnum;
|
|
|
|
|
|
|
|
Next = FirstFont;
|
|
|
|
FirstFont = this;
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSinglePicFont :: GetChar
|
|
|
|
//
|
|
|
|
// Returns the texture if code is 'a' or 'A', otherwise NULL.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
FTexture *FSinglePicFont::GetChar (int code, int *const width) const
|
|
|
|
{
|
|
|
|
*width = SpaceWidth;
|
|
|
|
if (code == 'a' || code == 'A')
|
|
|
|
{
|
|
|
|
return TexMan(PicNum);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSinglePicFont :: GetCharWidth
|
|
|
|
//
|
|
|
|
// Don't expect the text functions to work properly if I actually allowed
|
|
|
|
// the character width to vary depending on the animation frame.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
int FSinglePicFont::GetCharWidth (int code) const
|
|
|
|
{
|
|
|
|
return SpaceWidth;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: FFontChar1
|
|
|
|
//
|
|
|
|
// Used by fonts made from textures.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FFontChar1::FFontChar1 (FTexture *sourcelump)
|
|
|
|
: SourceRemap (NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
UseType = FTexture::TEX_FontChar;
|
2008-11-30 12:49:27 +00:00
|
|
|
BaseTexture = sourcelump;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
// now copy all the properties from the base texture
|
2008-11-30 12:49:27 +00:00
|
|
|
assert(BaseTexture != NULL);
|
|
|
|
CopySize(BaseTexture);
|
2006-02-24 04:48:15 +00:00
|
|
|
Pixels = NULL;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: GetPixels
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
const BYTE *FFontChar1::GetPixels ()
|
|
|
|
{
|
|
|
|
if (Pixels == NULL)
|
|
|
|
{
|
|
|
|
MakeTexture ();
|
|
|
|
}
|
|
|
|
return Pixels;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: MakeTexture
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FFontChar1::MakeTexture ()
|
|
|
|
{
|
|
|
|
// Make the texture as normal, then remap it so that all the colors
|
|
|
|
// are at the low end of the palette
|
|
|
|
Pixels = new BYTE[Width*Height];
|
|
|
|
const BYTE *pix = BaseTexture->GetPixels();
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
if (!SourceRemap)
|
|
|
|
{
|
|
|
|
memcpy(Pixels, pix, Width*Height);
|
|
|
|
}
|
|
|
|
else
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
for (int x = 0; x < Width*Height; ++x)
|
|
|
|
{
|
|
|
|
Pixels[x] = SourceRemap[pix[x]];
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: GetColumn
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
const BYTE *FFontChar1::GetColumn (unsigned int column, const Span **spans_out)
|
|
|
|
{
|
|
|
|
if (Pixels == NULL)
|
|
|
|
{
|
|
|
|
MakeTexture ();
|
|
|
|
}
|
|
|
|
|
|
|
|
BaseTexture->GetColumn(column, spans_out);
|
|
|
|
return Pixels + column*Height;
|
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: SetSourceRemap
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFontChar1::SetSourceRemap(const BYTE *sourceremap)
|
|
|
|
{
|
|
|
|
Unload();
|
|
|
|
SourceRemap = sourceremap;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: Unload
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FFontChar1::Unload ()
|
|
|
|
{
|
|
|
|
if (Pixels != NULL)
|
|
|
|
{
|
|
|
|
delete[] Pixels;
|
|
|
|
Pixels = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar1 :: ~FFontChar1
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
FFontChar1::~FFontChar1 ()
|
|
|
|
{
|
|
|
|
Unload ();
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: FFontChar2
|
|
|
|
//
|
|
|
|
// Used by FON1 and FON2 fonts.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FFontChar2::FFontChar2 (int sourcelump, int sourcepos, int width, int height, int leftofs, int topofs)
|
|
|
|
: SourceLump (sourcelump), SourcePos (sourcepos), Pixels (0), Spans (0), SourceRemap(NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
UseType = TEX_FontChar;
|
|
|
|
Width = width;
|
|
|
|
Height = height;
|
2010-02-26 05:34:30 +00:00
|
|
|
LeftOffset = leftofs;
|
|
|
|
TopOffset = topofs;
|
2006-02-24 04:48:15 +00:00
|
|
|
CalcBitSize ();
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: ~FFontChar2
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
FFontChar2::~FFontChar2 ()
|
|
|
|
{
|
|
|
|
Unload ();
|
|
|
|
if (Spans != NULL)
|
|
|
|
{
|
|
|
|
FreeSpans (Spans);
|
2006-05-04 06:14:52 +00:00
|
|
|
Spans = NULL;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: Unload
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FFontChar2::Unload ()
|
|
|
|
{
|
|
|
|
if (Pixels != NULL)
|
|
|
|
{
|
|
|
|
delete[] Pixels;
|
|
|
|
Pixels = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: GetPixels
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
const BYTE *FFontChar2::GetPixels ()
|
|
|
|
{
|
|
|
|
if (Pixels == NULL)
|
|
|
|
{
|
|
|
|
MakeTexture ();
|
|
|
|
}
|
|
|
|
return Pixels;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: GetColumn
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
const BYTE *FFontChar2::GetColumn (unsigned int column, const Span **spans_out)
|
|
|
|
{
|
|
|
|
if (Pixels == NULL)
|
|
|
|
{
|
|
|
|
MakeTexture ();
|
|
|
|
}
|
|
|
|
if (column >= Width)
|
|
|
|
{
|
|
|
|
column = WidthMask;
|
|
|
|
}
|
|
|
|
if (spans_out != NULL)
|
|
|
|
{
|
2008-04-14 12:10:45 +00:00
|
|
|
if (Spans == NULL)
|
|
|
|
{
|
|
|
|
Spans = CreateSpans (Pixels);
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
*spans_out = Spans[column];
|
|
|
|
}
|
|
|
|
return Pixels + column*Height;
|
|
|
|
}
|
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: SetSourceRemap
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFontChar2::SetSourceRemap(const BYTE *sourceremap)
|
|
|
|
{
|
|
|
|
Unload();
|
|
|
|
SourceRemap = sourceremap;
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFontChar2 :: MakeTexture
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
void FFontChar2::MakeTexture ()
|
|
|
|
{
|
|
|
|
FWadLump lump = Wads.OpenLumpNum (SourceLump);
|
|
|
|
int destSize = Width * Height;
|
|
|
|
BYTE max = 255;
|
2010-02-26 05:34:30 +00:00
|
|
|
bool rle = true;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
// This is to "fix" bad fonts
|
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
BYTE buff[16];
|
2006-02-24 04:48:15 +00:00
|
|
|
lump.Read (buff, 4);
|
|
|
|
if (buff[3] == '2')
|
|
|
|
{
|
|
|
|
lump.Read (buff, 7);
|
|
|
|
max = buff[6];
|
|
|
|
lump.Seek (SourcePos - 11, SEEK_CUR);
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
else if (buff[3] == 0x1A)
|
|
|
|
{
|
|
|
|
lump.Read(buff, 13);
|
|
|
|
max = buff[12] - 1;
|
|
|
|
lump.Seek (SourcePos - 17, SEEK_CUR);
|
|
|
|
rle = false;
|
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
else
|
|
|
|
{
|
|
|
|
lump.Seek (SourcePos - 4, SEEK_CUR);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
Pixels = new BYTE[destSize];
|
|
|
|
|
|
|
|
int runlen = 0, setlen = 0;
|
|
|
|
BYTE setval = 0; // Shut up, GCC!
|
|
|
|
BYTE *dest_p = Pixels;
|
|
|
|
int dest_adv = Height;
|
|
|
|
int dest_rew = destSize - 1;
|
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
if (rle)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
for (int y = Height; y != 0; --y)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
for (int x = Width; x != 0; )
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
if (runlen != 0)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
BYTE color;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
lump >> color;
|
|
|
|
color = MIN (color, max);
|
|
|
|
if (SourceRemap != NULL)
|
|
|
|
{
|
|
|
|
color = SourceRemap[color];
|
|
|
|
}
|
|
|
|
*dest_p = color;
|
|
|
|
dest_p += dest_adv;
|
|
|
|
x--;
|
|
|
|
runlen--;
|
|
|
|
}
|
|
|
|
else if (setlen != 0)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
*dest_p = setval;
|
|
|
|
dest_p += dest_adv;
|
|
|
|
x--;
|
|
|
|
setlen--;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
else
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
SBYTE code;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
lump >> code;
|
|
|
|
if (code >= 0)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
2010-02-26 05:34:30 +00:00
|
|
|
runlen = code + 1;
|
|
|
|
}
|
|
|
|
else if (code != -128)
|
|
|
|
{
|
|
|
|
BYTE color;
|
|
|
|
|
|
|
|
lump >> color;
|
|
|
|
setlen = (-code) + 1;
|
|
|
|
setval = MIN (color, max);
|
|
|
|
if (SourceRemap != NULL)
|
|
|
|
{
|
|
|
|
setval = SourceRemap[setval];
|
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
dest_p -= dest_rew;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
for (int y = Height; y != 0; --y)
|
|
|
|
{
|
|
|
|
for (int x = Width; x != 0; --x)
|
|
|
|
{
|
|
|
|
BYTE color;
|
|
|
|
lump >> color;
|
|
|
|
if (color > max)
|
|
|
|
{
|
|
|
|
color = max;
|
|
|
|
}
|
|
|
|
if (SourceRemap != NULL)
|
|
|
|
{
|
|
|
|
color = SourceRemap[color];
|
|
|
|
}
|
|
|
|
*dest_p = color;
|
|
|
|
dest_p += dest_adv;
|
|
|
|
}
|
|
|
|
dest_p -= dest_rew;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (destSize < 0)
|
|
|
|
{
|
|
|
|
char name[9];
|
|
|
|
Wads.GetLumpName (name, SourceLump);
|
|
|
|
name[8] = 0;
|
|
|
|
I_FatalError ("The font %s is corrupt", name);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSpecialFont :: FSpecialFont
|
|
|
|
//
|
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
FSpecialFont::FSpecialFont (const char *name, int first, int count, FTexture **lumplist, const bool *notranslate, int lump) : FFont(lump)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-06 20:59:52 +00:00
|
|
|
int i;
|
2008-11-30 12:49:27 +00:00
|
|
|
FTexture **charlumps;
|
2006-02-24 04:48:15 +00:00
|
|
|
int maxyoffs;
|
2008-09-04 14:09:32 +00:00
|
|
|
FTexture *pic;
|
2012-08-06 10:25:51 +00:00
|
|
|
|
|
|
|
memcpy(this->notranslate, notranslate, 256*sizeof(bool));
|
|
|
|
|
2008-11-30 01:37:11 +00:00
|
|
|
Name = copystring(name);
|
2006-02-24 04:48:15 +00:00
|
|
|
Chars = new CharData[count];
|
2008-11-30 12:49:27 +00:00
|
|
|
charlumps = new FTexture*[count];
|
2006-02-24 04:48:15 +00:00
|
|
|
PatchRemap = new BYTE[256];
|
|
|
|
FirstChar = first;
|
|
|
|
LastChar = first + count - 1;
|
|
|
|
FontHeight = 0;
|
|
|
|
GlobalKerning = false;
|
|
|
|
Next = FirstFont;
|
|
|
|
FirstFont = this;
|
|
|
|
|
|
|
|
maxyoffs = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++)
|
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
pic = charlumps[i] = lumplist[i];
|
|
|
|
if (pic != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
int height = pic->GetScaledHeight();
|
|
|
|
int yoffs = pic->GetScaledTopOffset();
|
|
|
|
|
|
|
|
if (yoffs > maxyoffs)
|
2008-09-04 14:09:32 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
maxyoffs = yoffs;
|
2008-09-04 14:09:32 +00:00
|
|
|
}
|
2008-11-30 12:49:27 +00:00
|
|
|
height += abs (yoffs);
|
|
|
|
if (height > FontHeight)
|
2008-09-04 14:09:32 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
FontHeight = height;
|
2008-09-04 14:09:32 +00:00
|
|
|
}
|
2012-08-06 10:25:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if (charlumps[i] != NULL)
|
|
|
|
{
|
|
|
|
Chars[i].Pic = new FFontChar1 (charlumps[i]);
|
|
|
|
Chars[i].XMove = Chars[i].Pic->GetScaledWidth();
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
Chars[i].Pic = NULL;
|
|
|
|
Chars[i].XMove = INT_MIN;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Special fonts normally don't have all characters so be careful here!
|
|
|
|
if ('N'-first >= 0 && 'N'-first < count && Chars['N' - first].Pic != NULL)
|
|
|
|
{
|
|
|
|
SpaceWidth = (Chars['N' - first].XMove + 1) / 2;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
SpaceWidth = 4;
|
|
|
|
}
|
|
|
|
|
|
|
|
FixXMoves();
|
|
|
|
|
|
|
|
LoadTranslations();
|
|
|
|
|
|
|
|
delete[] charlumps;
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FSpecialFont :: LoadTranslations
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FSpecialFont::LoadTranslations()
|
|
|
|
{
|
|
|
|
int count = LastChar - FirstChar + 1;
|
|
|
|
BYTE usedcolors[256], identity[256];
|
|
|
|
double *luminosity;
|
|
|
|
int TotalColors;
|
|
|
|
int i, j;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2012-08-06 10:25:51 +00:00
|
|
|
memset (usedcolors, 0, 256);
|
|
|
|
for (i = 0; i < count; i++)
|
|
|
|
{
|
|
|
|
FFontChar1 *pic = static_cast<FFontChar1 *>(Chars[i].Pic);
|
|
|
|
if(pic)
|
|
|
|
{
|
|
|
|
pic->SetSourceRemap(NULL); // Force the FFontChar1 to return the same pixels as the base texture
|
2008-11-30 12:49:27 +00:00
|
|
|
RecordTextureColors (pic, usedcolors);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// exclude the non-translated colors from the translation calculation
|
2014-03-01 21:27:40 +00:00
|
|
|
for (i = 0; i < 256; i++)
|
|
|
|
if (notranslate[i])
|
|
|
|
usedcolors[i] = false;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
TotalColors = ActiveColors = SimpleTranslation (usedcolors, PatchRemap, identity, &luminosity);
|
|
|
|
|
|
|
|
// Map all untranslated colors into the table of used colors
|
2014-03-01 21:27:40 +00:00
|
|
|
for (i = 0; i < 256; i++)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2014-03-01 21:27:40 +00:00
|
|
|
if (notranslate[i])
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2014-03-01 21:27:40 +00:00
|
|
|
PatchRemap[i] = TotalColors;
|
|
|
|
identity[TotalColors] = i;
|
|
|
|
TotalColors++;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++)
|
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
if(Chars[i].Pic)
|
|
|
|
static_cast<FFontChar1 *>(Chars[i].Pic)->SetSourceRemap(PatchRemap);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
BuildTranslations (luminosity, identity, &TranslationParms[0][0], TotalColors, NULL);
|
2006-02-24 04:48:15 +00:00
|
|
|
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
// add the untranslated colors to the Ranges tables
|
2006-02-24 04:48:15 +00:00
|
|
|
if (ActiveColors < TotalColors)
|
|
|
|
{
|
2007-12-23 16:53:39 +00:00
|
|
|
for (i = 0; i < NumTextColors; i++)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
FRemapTable *remap = &Ranges[i];
|
|
|
|
for (j = ActiveColors; j < TotalColors; ++j)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Discovered that Shader Model 1.4 clamps my constants, so I can't use
palettes smaller than 256 entries with the shader I wrote for it. Is there
a list of gotchas like this listed some where? I'd really like to see it.
Well, when compiled with SM2.0, the PalTex shader seems to be every-so-
slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a
minor win for cards that support it.
- Fixed: ST_Endoom() failed to free the bitmap it used.
- Added the DTA_ColorOverlay attribute to blend a color with the texture
being drawn. For software, this (currently) only works with black. For
hardware, it works with any color. The motiviation for this was so I could
rewrite the status bar calls that passed DIM_MAP to DTA_Translation to
draw darker icons into something that didn't require making a whole new
remap table.
- After having an "OMG! How could I have been so stupid?" moment, I have
removed the off-by-one check from D3DFB. I had thought the off-by-one error
was caused by rounding errors by the shader hardware. Not so. Rather, I
wasn't sampling what I thought I was sampling. A texture that uses palette
index 255 passes the value 1.0 to the shader. The shader needs to adjust the
range of its palette indexes, or it will end up trying to read color 256
from the palette texture when it should be reading color 255. Doh!
- The TranslationToTable() function has been added to map from translation
numbers used by actors to the tables those numbers represent. This function
performs validation for the input and returns NULL if the input value
is invalid.
- Major changes to the way translation tables work: No longer are they each a
256-byte array. Instead, the FRemapTable structure is used to represent each
one. It includes a remap array for the software renderer, a palette array
for a hardware renderer, and a native texture pointer for D3DFB. The
translationtables array itself is now an array of TArrays that point to the
real tables. The DTA_Translation attribute must also be passed a pointer
to a FRemapTable, not a byte array as previously.
- Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly
for D3DFB's 2D mode. Before, any fullscreen graphics (like help images)
covered it up.
SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
|
|
|
remap->Remap[j] = identity[j];
|
2007-12-26 21:37:00 +00:00
|
|
|
remap->Palette[j] = GPalette.BaseColors[identity[j]];
|
2009-02-21 11:30:42 +00:00
|
|
|
remap->Palette[j].a = 0xff;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
ActiveColors = TotalColors;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
delete[] luminosity;
|
|
|
|
}
|
|
|
|
|
2010-02-26 05:34:30 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// FFont :: FixXMoves
|
|
|
|
//
|
|
|
|
// If a font has gaps in its characters, set the missing characters'
|
2013-05-17 02:35:35 +00:00
|
|
|
// XMoves to either SpaceWidth or the unaccented or uppercase variant's
|
|
|
|
// XMove. Missing XMoves must be initialized with INT_MIN beforehand.
|
2010-02-26 05:34:30 +00:00
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void FFont::FixXMoves()
|
|
|
|
{
|
|
|
|
for (int i = 0; i <= LastChar - FirstChar; ++i)
|
|
|
|
{
|
|
|
|
if (Chars[i].XMove == INT_MIN)
|
|
|
|
{
|
2013-05-17 02:35:35 +00:00
|
|
|
// Try an uppercase character.
|
|
|
|
if (myislower(i + FirstChar))
|
2010-02-26 05:34:30 +00:00
|
|
|
{
|
|
|
|
int upper = i - 32;
|
|
|
|
if (upper >= 0)
|
|
|
|
{
|
|
|
|
Chars[i].XMove = Chars[upper].XMove;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
2013-05-17 02:35:35 +00:00
|
|
|
// Try an unnaccented character.
|
|
|
|
int noaccent = stripaccent(i + FirstChar);
|
|
|
|
if (noaccent != i + FirstChar)
|
|
|
|
{
|
|
|
|
noaccent -= FirstChar;
|
|
|
|
if (noaccent >= 0)
|
|
|
|
{
|
|
|
|
Chars[i].XMove = Chars[noaccent].XMove;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
2010-02-26 05:34:30 +00:00
|
|
|
Chars[i].XMove = SpaceWidth;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-02-24 04:48:15 +00:00
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_InitCustomFonts
|
|
|
|
//
|
2006-02-24 04:48:15 +00:00
|
|
|
// Initialize a list of custom multipatch fonts
|
|
|
|
//
|
2006-08-30 02:38:39 +00:00
|
|
|
//==========================================================================
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
void V_InitCustomFonts()
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
FScanner sc;
|
2008-11-30 12:49:27 +00:00
|
|
|
FTexture *lumplist[256];
|
2006-02-24 04:48:15 +00:00
|
|
|
bool notranslate[256];
|
2008-09-04 14:09:32 +00:00
|
|
|
FString namebuffer, templatebuf;
|
2006-02-24 04:48:15 +00:00
|
|
|
int i;
|
2006-04-23 09:02:19 +00:00
|
|
|
int llump,lastlump=0;
|
2006-02-24 04:48:15 +00:00
|
|
|
int format;
|
|
|
|
int start;
|
|
|
|
int first;
|
|
|
|
int count;
|
2012-08-07 08:11:56 +00:00
|
|
|
int spacewidth;
|
2010-10-12 08:43:15 +00:00
|
|
|
char cursor = '_';
|
2006-02-24 04:48:15 +00:00
|
|
|
|
|
|
|
while ((llump = Wads.FindLump ("FONTDEFS", &lastlump)) != -1)
|
|
|
|
{
|
2008-04-03 03:19:21 +00:00
|
|
|
sc.OpenLumpNum(llump);
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
while (sc.GetString())
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
memset (lumplist, 0, sizeof(lumplist));
|
2006-02-24 04:48:15 +00:00
|
|
|
memset (notranslate, 0, sizeof(notranslate));
|
2008-09-04 14:09:32 +00:00
|
|
|
namebuffer = sc.String;
|
2006-02-24 04:48:15 +00:00
|
|
|
format = 0;
|
|
|
|
start = 33;
|
|
|
|
first = 33;
|
|
|
|
count = 223;
|
2012-08-07 08:11:56 +00:00
|
|
|
spacewidth = -1;
|
2006-02-24 04:48:15 +00:00
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetStringName ("{");
|
|
|
|
while (!sc.CheckString ("}"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetString();
|
|
|
|
if (sc.Compare ("TEMPLATE"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (format == 2) goto wrong;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetString();
|
2008-09-04 14:09:32 +00:00
|
|
|
templatebuf = sc.String;
|
2006-02-24 04:48:15 +00:00
|
|
|
format = 1;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (sc.Compare ("BASE"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (format == 2) goto wrong;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetNumber();
|
|
|
|
start = sc.Number;
|
2006-02-24 04:48:15 +00:00
|
|
|
format = 1;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (sc.Compare ("FIRST"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (format == 2) goto wrong;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetNumber();
|
|
|
|
first = sc.Number;
|
2006-02-24 04:48:15 +00:00
|
|
|
format = 1;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (sc.Compare ("COUNT"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (format == 2) goto wrong;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetNumber();
|
|
|
|
count = sc.Number;
|
2006-02-24 04:48:15 +00:00
|
|
|
format = 1;
|
|
|
|
}
|
2010-10-12 08:43:15 +00:00
|
|
|
else if (sc.Compare ("CURSOR"))
|
|
|
|
{
|
|
|
|
sc.MustGetString();
|
|
|
|
cursor = sc.String[0];
|
|
|
|
}
|
2012-08-07 08:11:56 +00:00
|
|
|
else if (sc.Compare ("SPACEWIDTH"))
|
|
|
|
{
|
|
|
|
if (format == 2) goto wrong;
|
|
|
|
sc.MustGetNumber();
|
|
|
|
spacewidth = sc.Number;
|
|
|
|
format = 1;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (sc.Compare ("NOTRANSLATION"))
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
if (format == 1) goto wrong;
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
while (sc.CheckNumber() && !sc.Crossed)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.Number >= 0 && sc.Number < 256)
|
|
|
|
notranslate[sc.Number] = true;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
2008-11-27 17:43:36 +00:00
|
|
|
format = 2;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (format == 1) goto wrong;
|
2008-11-30 12:49:27 +00:00
|
|
|
FTexture **p = &lumplist[*(unsigned char*)sc.String];
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetString();
|
2008-11-30 12:49:27 +00:00
|
|
|
FTextureID texid = TexMan.CheckForTexture(sc.String, FTexture::TEX_MiscPatch);
|
|
|
|
if (texid.Exists())
|
|
|
|
{
|
|
|
|
*p = TexMan[texid];
|
|
|
|
}
|
|
|
|
else if (Wads.GetLumpFile(sc.LumpNum) >= Wads.IWAD_FILENUM)
|
2008-11-30 01:37:11 +00:00
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
// Print a message only if this isn't in zdoom.pk3
|
|
|
|
sc.ScriptMessage("%s: Unable to find texture in font definition for %s", sc.String, namebuffer.GetChars());
|
2008-11-30 01:37:11 +00:00
|
|
|
}
|
|
|
|
format = 2;
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (format == 1)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-07 08:11:56 +00:00
|
|
|
FFont *fnt = new FFont (namebuffer, templatebuf, first, count, start, llump, spacewidth);
|
2010-10-12 08:43:15 +00:00
|
|
|
fnt->SetCursor(cursor);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (format == 2)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
for (i = 0; i < 256; i++)
|
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
if (lumplist[i] != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
first = i;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (i = 255; i >= 0; i--)
|
|
|
|
{
|
2008-11-30 12:49:27 +00:00
|
|
|
if (lumplist[i] != NULL)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
|
|
|
count = i - first + 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2008-11-27 17:43:36 +00:00
|
|
|
if (count > 0)
|
2006-02-24 04:48:15 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
FFont *fnt = new FSpecialFont (namebuffer, first, count, &lumplist[first], notranslate, llump);
|
2010-10-12 08:43:15 +00:00
|
|
|
fnt->SetCursor(cursor);
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else goto wrong;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.Close();
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
|
|
|
return;
|
|
|
|
|
|
|
|
wrong:
|
2008-09-17 20:24:08 +00:00
|
|
|
sc.ScriptError ("Invalid combination of properties in font '%s'", namebuffer.GetChars());
|
2006-02-24 04:48:15 +00:00
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_InitFontColors
|
|
|
|
//
|
|
|
|
// Reads the list of color translation definitions into memory.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void V_InitFontColors ()
|
|
|
|
{
|
|
|
|
TArray<FName> names;
|
|
|
|
int lump, lastlump = 0;
|
2011-03-29 05:20:33 +00:00
|
|
|
TranslationParm tparm = { 0, 0, {0}, {0} }; // Silence GCC (for real with -Wextra )
|
2006-08-30 02:38:39 +00:00
|
|
|
TArray<TranslationParm> parms;
|
|
|
|
TArray<TempParmInfo> parminfo;
|
|
|
|
TArray<TempColorInfo> colorinfo;
|
|
|
|
int c, parmchoice;
|
|
|
|
TempParmInfo info;
|
|
|
|
TempColorInfo cinfo;
|
2007-01-09 04:40:58 +00:00
|
|
|
PalEntry logcolor;
|
2006-08-30 02:38:39 +00:00
|
|
|
unsigned int i, j;
|
|
|
|
int k, index;
|
|
|
|
|
|
|
|
info.Index = -1;
|
|
|
|
|
2010-12-15 00:09:31 +00:00
|
|
|
TranslationParms[0].Clear();
|
|
|
|
TranslationParms[1].Clear();
|
|
|
|
TranslationLookup.Clear();
|
|
|
|
TranslationColors.Clear();
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
while ((lump = Wads.FindLump ("TEXTCOLO", &lastlump)) != -1)
|
|
|
|
{
|
2010-09-14 17:28:18 +00:00
|
|
|
if (gameinfo.flags & GI_NOTEXTCOLOR)
|
|
|
|
{
|
|
|
|
// Chex3 contains a bad TEXTCOLO lump, probably to force all text to be green.
|
|
|
|
// This renders the Gray, Gold, Red and Yellow color range inoperable, some of
|
|
|
|
// which are used by the menu. So we have no choice but to skip this lump so that
|
|
|
|
// all colors work properly.
|
|
|
|
// The text colors should be the end user's choice anyway.
|
|
|
|
if (Wads.GetLumpFile(lump) == 1) continue;
|
|
|
|
}
|
2008-04-03 03:19:21 +00:00
|
|
|
FScanner sc(lump);
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
while (sc.GetString())
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
|
|
|
names.Clear();
|
|
|
|
|
2007-01-09 04:40:58 +00:00
|
|
|
logcolor = DEFAULT_LOG_COLOR;
|
|
|
|
|
2006-08-30 02:38:39 +00:00
|
|
|
// Everything until the '{' is considered a valid name for the
|
|
|
|
// color range.
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
names.Push (sc.String);
|
|
|
|
while (sc.MustGetString(), !sc.Compare ("{"))
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
|
|
|
if (names[0] == NAME_Untranslated)
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The \"untranslated\" color may not have any other names");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
names.Push (sc.String);
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
parmchoice = 0;
|
|
|
|
info.StartParm[0] = parms.Size();
|
|
|
|
info.StartParm[1] = 0;
|
|
|
|
info.ParmLen[1] = info.ParmLen[0] = 0;
|
|
|
|
tparm.RangeEnd = tparm.RangeStart = -1;
|
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
while (sc.MustGetString(), !sc.Compare ("}"))
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.Compare ("Console:"))
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
|
|
|
if (parmchoice == 1)
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("Each color may only have one set of console ranges");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
parmchoice = 1;
|
|
|
|
info.StartParm[1] = parms.Size();
|
|
|
|
info.ParmLen[0] = info.StartParm[1] - info.StartParm[0];
|
|
|
|
tparm.RangeEnd = tparm.RangeStart = -1;
|
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
else if (sc.Compare ("Flat:"))
|
2007-01-09 04:40:58 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetString();
|
|
|
|
logcolor = V_GetColor (NULL, sc.String);
|
2007-01-09 04:40:58 +00:00
|
|
|
}
|
2006-08-30 02:38:39 +00:00
|
|
|
else
|
|
|
|
{
|
|
|
|
// Get first color
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
c = V_GetColor (NULL, sc.String);
|
2006-08-30 02:38:39 +00:00
|
|
|
tparm.Start[0] = RPART(c);
|
|
|
|
tparm.Start[1] = GPART(c);
|
|
|
|
tparm.Start[2] = BPART(c);
|
|
|
|
|
|
|
|
// Get second color
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetString();
|
|
|
|
c = V_GetColor (NULL, sc.String);
|
2006-08-30 02:38:39 +00:00
|
|
|
tparm.End[0] = RPART(c);
|
|
|
|
tparm.End[1] = GPART(c);
|
|
|
|
tparm.End[2] = BPART(c);
|
|
|
|
|
|
|
|
// Check for range specifier
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.CheckNumber())
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (tparm.RangeStart == -1 && sc.Number != 0)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The first color range must start at position 0");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.Number < 0 || sc.Number > 256)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The color range must be within positions [0,256]");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.Number <= tparm.RangeEnd)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The color range must not start before the previous one ends");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
tparm.RangeStart = sc.Number;
|
2006-08-30 02:38:39 +00:00
|
|
|
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.MustGetNumber();
|
|
|
|
if (sc.Number < 0 || sc.Number > 256)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The color range must be within positions [0,256]");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
if (sc.Number <= tparm.RangeStart)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The color range end position must be larger than the start position");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
tparm.RangeEnd = sc.Number;
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
tparm.RangeStart = tparm.RangeEnd + 1;
|
|
|
|
tparm.RangeEnd = 256;
|
|
|
|
if (tparm.RangeStart >= tparm.RangeEnd)
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The color has too many ranges");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
parms.Push (tparm);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
info.ParmLen[parmchoice] = parms.Size() - info.StartParm[parmchoice];
|
|
|
|
if (info.ParmLen[0] == 0)
|
|
|
|
{
|
|
|
|
if (names[0] != NAME_Untranslated)
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("There must be at least one normal range for a color");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (names[0] == NAME_Untranslated)
|
|
|
|
{
|
- Updated lempar.c to v1.31.
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
2008-01-25 23:57:44 +00:00
|
|
|
sc.ScriptError ("The \"untranslated\" color must be left undefined");
|
2006-08-30 02:38:39 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
if (info.ParmLen[1] == 0 && names[0] != NAME_Untranslated)
|
|
|
|
{ // If a console translation is unspecified, make it white, since the console
|
|
|
|
// font has no color information stored with it.
|
|
|
|
tparm.RangeStart = 0;
|
|
|
|
tparm.RangeEnd = 256;
|
|
|
|
tparm.Start[2] = tparm.Start[1] = tparm.Start[0] = 0;
|
|
|
|
tparm.End[2] = tparm.End[1] = tparm.End[0] = 255;
|
|
|
|
info.StartParm[1] = parms.Push (tparm);
|
|
|
|
info.ParmLen[1] = 1;
|
|
|
|
}
|
|
|
|
cinfo.ParmInfo = parminfo.Push (info);
|
|
|
|
// Record this color information for each name it goes by
|
|
|
|
for (i = 0; i < names.Size(); ++i)
|
|
|
|
{
|
|
|
|
// Redefine duplicates in-place
|
|
|
|
for (j = 0; j < colorinfo.Size(); ++j)
|
|
|
|
{
|
|
|
|
if (colorinfo[j].Name == names[i])
|
|
|
|
{
|
|
|
|
colorinfo[j].ParmInfo = cinfo.ParmInfo;
|
2007-01-09 04:40:58 +00:00
|
|
|
colorinfo[j].LogColor = logcolor;
|
2006-08-30 02:38:39 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (j == colorinfo.Size())
|
|
|
|
{
|
|
|
|
cinfo.Name = names[i];
|
2007-01-09 04:40:58 +00:00
|
|
|
cinfo.LogColor = logcolor;
|
2006-08-30 02:38:39 +00:00
|
|
|
colorinfo.Push (cinfo);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
// Make permananent copies of all the color information we found.
|
|
|
|
for (i = 0, index = 0; i < colorinfo.Size(); ++i)
|
|
|
|
{
|
|
|
|
TranslationMap tmap;
|
|
|
|
TempParmInfo *pinfo;
|
|
|
|
|
|
|
|
tmap.Name = colorinfo[i].Name;
|
|
|
|
pinfo = &parminfo[colorinfo[i].ParmInfo];
|
|
|
|
if (pinfo->Index < 0)
|
|
|
|
{
|
|
|
|
// Write out the set of remappings for this color.
|
|
|
|
for (k = 0; k < 2; ++k)
|
|
|
|
{
|
|
|
|
for (j = 0; j < pinfo->ParmLen[k]; ++j)
|
|
|
|
{
|
|
|
|
TranslationParms[k].Push (parms[pinfo->StartParm[k] + j]);
|
|
|
|
}
|
|
|
|
}
|
2007-01-09 04:40:58 +00:00
|
|
|
TranslationColors.Push (colorinfo[i].LogColor);
|
2006-08-30 02:38:39 +00:00
|
|
|
pinfo->Index = index++;
|
|
|
|
}
|
|
|
|
tmap.Number = pinfo->Index;
|
|
|
|
TranslationLookup.Push (tmap);
|
|
|
|
}
|
|
|
|
// Leave a terminating marker at the ends of the lists.
|
|
|
|
tparm.RangeStart = -1;
|
|
|
|
TranslationParms[0].Push (tparm);
|
|
|
|
TranslationParms[1].Push (tparm);
|
|
|
|
// Sort the translation lookups for fast binary searching.
|
|
|
|
qsort (&TranslationLookup[0], TranslationLookup.Size(), sizeof(TranslationLookup[0]), TranslationMapCompare);
|
|
|
|
|
|
|
|
NumTextColors = index;
|
|
|
|
assert (NumTextColors >= NUM_TEXT_COLORS);
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// TranslationMapCompare
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
static int STACK_ARGS TranslationMapCompare (const void *a, const void *b)
|
|
|
|
{
|
|
|
|
return int(((const TranslationMap *)a)->Name) - int(((const TranslationMap *)b)->Name);
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_FindFontColor
|
|
|
|
//
|
|
|
|
// Returns the color number for a particular named color range.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
2006-08-31 00:16:12 +00:00
|
|
|
EColorRange V_FindFontColor (FName name)
|
2006-08-30 02:38:39 +00:00
|
|
|
{
|
2006-08-31 00:16:12 +00:00
|
|
|
int min = 0, max = TranslationLookup.Size() - 1;
|
2006-08-30 02:38:39 +00:00
|
|
|
|
|
|
|
while (min <= max)
|
|
|
|
{
|
|
|
|
unsigned int mid = (min + max) / 2;
|
|
|
|
const TranslationMap *probe = &TranslationLookup[mid];
|
|
|
|
if (probe->Name == name)
|
|
|
|
{
|
|
|
|
return EColorRange(probe->Number);
|
|
|
|
}
|
|
|
|
else if (probe->Name < name)
|
|
|
|
{
|
|
|
|
min = mid + 1;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
max = mid - 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return CR_UNTRANSLATED;
|
|
|
|
}
|
2006-10-06 21:07:58 +00:00
|
|
|
|
2007-01-09 04:40:58 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_LogColorFromColorRange
|
|
|
|
//
|
|
|
|
// Returns the color to use for text in the startup/error log window.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
PalEntry V_LogColorFromColorRange (EColorRange range)
|
|
|
|
{
|
2007-01-16 01:53:43 +00:00
|
|
|
if ((unsigned int)range >= TranslationColors.Size())
|
2007-01-09 04:40:58 +00:00
|
|
|
{ // Return default color
|
|
|
|
return DEFAULT_LOG_COLOR;
|
|
|
|
}
|
|
|
|
return TranslationColors[range];
|
|
|
|
}
|
|
|
|
|
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_ParseFontColor
|
|
|
|
//
|
|
|
|
// Given a pointer to a color identifier (presumably just after a color
|
|
|
|
// escape character), return the color it identifies and advances
|
|
|
|
// color_value to just past it.
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
EColorRange V_ParseFontColor (const BYTE *&color_value, int normalcolor, int boldcolor)
|
|
|
|
{
|
|
|
|
const BYTE *ch = color_value;
|
|
|
|
int newcolor = *ch++;
|
|
|
|
|
2009-02-20 00:53:25 +00:00
|
|
|
if (newcolor == '-') // Normal
|
2007-01-09 04:40:58 +00:00
|
|
|
{
|
|
|
|
newcolor = normalcolor;
|
|
|
|
}
|
|
|
|
else if (newcolor == '+') // Bold
|
|
|
|
{
|
|
|
|
newcolor = boldcolor;
|
|
|
|
}
|
2009-02-20 00:53:25 +00:00
|
|
|
else if (newcolor == '!') // Team chat
|
|
|
|
{
|
|
|
|
newcolor = PrintColors[PRINT_TEAMCHAT];
|
|
|
|
}
|
|
|
|
else if (newcolor == '*') // Chat
|
|
|
|
{
|
|
|
|
newcolor = PrintColors[PRINT_CHAT];
|
|
|
|
}
|
2007-01-09 04:40:58 +00:00
|
|
|
else if (newcolor == '[') // Named
|
|
|
|
{
|
|
|
|
const BYTE *namestart = ch;
|
|
|
|
while (*ch != ']' && *ch != '\0')
|
|
|
|
{
|
|
|
|
ch++;
|
|
|
|
}
|
|
|
|
FName rangename((const char *)namestart, int(ch - namestart), true);
|
|
|
|
if (*ch != '\0')
|
|
|
|
{
|
|
|
|
ch++;
|
|
|
|
}
|
|
|
|
newcolor = V_FindFontColor (rangename);
|
|
|
|
}
|
|
|
|
else if (newcolor >= 'A' && newcolor < NUM_TEXT_COLORS + 'A') // Standard, uppercase
|
|
|
|
{
|
|
|
|
newcolor -= 'A';
|
|
|
|
}
|
|
|
|
else if (newcolor >= 'a' && newcolor < NUM_TEXT_COLORS + 'a') // Standard, lowercase
|
|
|
|
{
|
|
|
|
newcolor -= 'a';
|
|
|
|
}
|
|
|
|
else // Incomplete!
|
|
|
|
{
|
|
|
|
color_value = ch - (*ch == '\0');
|
|
|
|
return CR_UNDEFINED;
|
|
|
|
}
|
|
|
|
color_value = ch;
|
|
|
|
return EColorRange(newcolor);
|
|
|
|
}
|
|
|
|
|
2006-10-06 21:07:58 +00:00
|
|
|
//==========================================================================
|
|
|
|
//
|
|
|
|
// V_InitFonts
|
|
|
|
//
|
|
|
|
//==========================================================================
|
|
|
|
|
|
|
|
void V_InitFonts()
|
|
|
|
{
|
2007-05-26 10:50:32 +00:00
|
|
|
V_InitCustomFonts ();
|
2006-10-06 21:07:58 +00:00
|
|
|
|
|
|
|
// load the heads-up font
|
2007-05-26 10:50:32 +00:00
|
|
|
if (!(SmallFont = FFont::FindFont("SmallFont")))
|
2006-10-06 21:07:58 +00:00
|
|
|
{
|
2008-11-30 01:37:11 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
if ((i = Wads.CheckNumForName("SMALLFNT")) >= 0)
|
|
|
|
{
|
|
|
|
SmallFont = new FSingleLumpFont("SmallFont", i);
|
|
|
|
}
|
|
|
|
else if (Wads.CheckNumForName ("FONTA_S") >= 0)
|
2007-05-26 10:50:32 +00:00
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
SmallFont = new FFont ("SmallFont", "FONTA%02u", HU_FONTSTART, HU_FONTSIZE, 1, -1);
|
2010-10-12 08:43:15 +00:00
|
|
|
SmallFont->SetCursor('[');
|
2007-05-26 10:50:32 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
SmallFont = new FFont ("SmallFont", "STCFN%.3d", HU_FONTSTART, HU_FONTSIZE, HU_FONTSTART, -1);
|
2007-05-26 10:50:32 +00:00
|
|
|
}
|
2006-10-06 21:07:58 +00:00
|
|
|
}
|
2008-11-30 01:37:11 +00:00
|
|
|
if (!(SmallFont2 = FFont::FindFont("SmallFont2"))) // Only used by Strife
|
2006-10-06 21:07:58 +00:00
|
|
|
{
|
2007-05-26 10:50:32 +00:00
|
|
|
if (Wads.CheckNumForName ("STBFN033", ns_graphics) >= 0)
|
|
|
|
{
|
2012-08-06 10:25:51 +00:00
|
|
|
SmallFont2 = new FFont ("SmallFont2", "STBFN%.3d", HU_FONTSTART, HU_FONTSIZE, HU_FONTSTART, -1);
|
2007-05-26 10:50:32 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
SmallFont2 = SmallFont;
|
|
|
|
}
|
2006-10-06 21:07:58 +00:00
|
|
|
}
|
2008-11-27 17:43:36 +00:00
|
|
|
if (!(BigFont = FFont::FindFont("BigFont")))
|
2006-10-06 21:07:58 +00:00
|
|
|
{
|
2013-06-01 02:04:44 +00:00
|
|
|
int lump = Wads.CheckNumForName("BIGFONT");
|
|
|
|
if (lump >= 0)
|
|
|
|
{
|
|
|
|
BigFont = new FSingleLumpFont("BigFont", lump);
|
|
|
|
}
|
|
|
|
else if (gameinfo.gametype & GAME_DoomChex)
|
2007-05-26 10:50:32 +00:00
|
|
|
{
|
|
|
|
BigFont = new FSingleLumpFont ("BigFont", Wads.GetNumForName ("DBIGFONT"));
|
|
|
|
}
|
|
|
|
else if (gameinfo.gametype == GAME_Strife)
|
|
|
|
{
|
|
|
|
BigFont = new FSingleLumpFont ("BigFont", Wads.GetNumForName ("SBIGFONT"));
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2013-06-01 02:04:44 +00:00
|
|
|
lump = Wads.CheckNumForName("HBIGFONT");
|
|
|
|
if (lump >= 0)
|
|
|
|
{
|
|
|
|
BigFont = new FSingleLumpFont("BigFont", lump);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
BigFont = new FFont ("BigFont", "FONTB%02u", HU_FONTSTART, HU_FONTSIZE, 1, -1);
|
|
|
|
}
|
2007-05-26 10:50:32 +00:00
|
|
|
}
|
2006-10-06 21:07:58 +00:00
|
|
|
}
|
2008-11-27 17:43:36 +00:00
|
|
|
if (!(ConFont = FFont::FindFont("ConsoleFont")))
|
2006-10-06 21:07:58 +00:00
|
|
|
{
|
2007-05-26 10:50:32 +00:00
|
|
|
ConFont = new FSingleLumpFont ("ConsoleFont", Wads.GetNumForName ("CONFONT"));
|
2006-10-06 21:07:58 +00:00
|
|
|
}
|
2008-11-27 17:43:36 +00:00
|
|
|
if (!(IntermissionFont = FFont::FindFont("IntermissionFont")))
|
|
|
|
{
|
|
|
|
if (gameinfo.gametype & GAME_DoomChex)
|
|
|
|
{
|
|
|
|
IntermissionFont = FFont::FindFont("IntermissionFont_Doom");
|
|
|
|
}
|
|
|
|
if (IntermissionFont == NULL)
|
|
|
|
{
|
|
|
|
IntermissionFont = BigFont;
|
|
|
|
}
|
|
|
|
}
|
2007-04-29 12:07:27 +00:00
|
|
|
}
|
2010-12-15 00:09:31 +00:00
|
|
|
|
|
|
|
void V_ClearFonts()
|
|
|
|
{
|
|
|
|
while (FFont::FirstFont != NULL)
|
|
|
|
{
|
|
|
|
delete FFont::FirstFont;
|
|
|
|
}
|
|
|
|
FFont::FirstFont = NULL;
|
|
|
|
SmallFont = SmallFont2 = BigFont = ConFont = IntermissionFont = NULL;
|
2012-08-06 10:25:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void V_RetranslateFonts()
|
|
|
|
{
|
|
|
|
FFont *font = FFont::FirstFont;
|
|
|
|
while(font)
|
|
|
|
{
|
|
|
|
font->LoadTranslations();
|
|
|
|
font = font->Next;
|
|
|
|
}
|
|
|
|
}
|