gzdoom-gles/src/r_things.cpp

2628 lines
66 KiB
C++
Raw Normal View History

// Emacs style mode select -*- C++ -*-
//-----------------------------------------------------------------------------
//
// $Id:$
//
// Copyright (C) 1993-1996 by id Software, Inc.
//
// This source is available for distribution and/or modification
// only under the terms of the DOOM Source Code License as
// published by id Software. All rights reserved.
//
// The source is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// FITNESS FOR A PARTICULAR PURPOSE. See the DOOM Source Code License
// for more details.
//
// $Log:$
//
// DESCRIPTION:
// Refresh of things, i.e. objects represented by sprites.
//
//-----------------------------------------------------------------------------
#include <stdio.h>
#include <stdlib.h>
#include <algorithm>
#include "templates.h"
#include "doomdef.h"
#include "m_swap.h"
#include "m_argv.h"
#include "i_system.h"
#include "w_wad.h"
#include "r_local.h"
#include "c_console.h"
#include "c_cvars.h"
#include "c_dispatch.h"
#include "doomstat.h"
#include "v_video.h"
#include "sc_man.h"
#include "s_sound.h"
#include "sbar.h"
#include "gi.h"
#include "r_sky.h"
#include "cmdlib.h"
#include "g_level.h"
#include "d_net.h"
#include "colormatcher.h"
2008-09-15 00:47:31 +00:00
#include "d_netinf.h"
#include "r_bsp.h"
#include "r_plane.h"
#include "r_segs.h"
#include "v_palette.h"
#include "r_translate.h"
extern fixed_t globaluclip, globaldclip;
#define MINZ (2048*4)
#define BASEYCENTER (100)
EXTERN_CVAR (Bool, st_scale)
CVAR (Bool, r_drawfuzz, true, CVAR_ARCHIVE)
//
// Sprite rotation 0 is facing the viewer,
// rotation 1 is one angle turn CLOCKWISE around the axis.
// This is not the same as the angle,
// which increases counter clockwise (protractor).
//
fixed_t pspritexscale;
fixed_t pspriteyscale;
fixed_t pspritexiscale;
fixed_t sky1scale; // [RH] Sky 1 scale factor
fixed_t sky2scale; // [RH] Sky 2 scale factor
vissprite_t *VisPSprites[NUMPSPRITES];
int VisPSpritesX1[NUMPSPRITES];
FDynamicColormap *VisPSpritesBaseColormap[NUMPSPRITES];
static int spriteshade;
TArray<WORD> ParticlesInSubsec;
// constant arrays
// used for psprite clipping and initializing clipping
short zeroarray[MAXWIDTH];
short screenheightarray[MAXWIDTH];
#define MAX_SPRITE_FRAMES 29 // [RH] Macro-ized as in BOOM.
CVAR (Bool, r_drawplayersprites, true, 0) // [RH] Draw player sprites?
//
// INITIALIZATION FUNCTIONS
//
// variables used to look up
// and range check thing_t sprites patches
TArray<spritedef_t> sprites;
TArray<spriteframe_t> SpriteFrames;
DWORD NumStdSprites; // The first x sprites that don't belong to skins.
struct spriteframewithrotate : public spriteframe_t
{
int rotate;
}
sprtemp[MAX_SPRITE_FRAMES];
int maxframe;
char* spritename;
// [RH] skin globals
FPlayerSkin *skins;
size_t numskins;
BYTE OtherGameSkinRemap[256];
- 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
PalEntry OtherGameSkinPalette[256];
// [RH] particle globals
WORD NumParticles;
WORD ActiveParticles;
WORD InactiveParticles;
particle_t *Particles;
CVAR (Bool, r_particles, true, 0);
//
// R_InstallSpriteLump
// Local function for R_InitSprites.
//
// [RH] Removed checks for coexistance of rotation 0 with other
// rotations and made it look more like BOOM's version.
//
static void R_InstallSpriteLump (FTextureID lump, unsigned frame, char rot, bool flipped)
{
unsigned rotation;
if (rot >= '0' && rot <= '9')
{
rotation = rot - '0';
}
else if (rot >= 'A')
{
rotation = rot - 'A' + 10;
}
else
{
rotation = 17;
}
if (frame >= MAX_SPRITE_FRAMES || rotation > 16)
I_FatalError ("R_InstallSpriteLump: Bad frame characters in lump %s", TexMan[lump]->Name);
if ((int)frame > maxframe)
maxframe = frame;
if (rotation == 0)
{
// the lump should be used for all rotations
// false=0, true=1, but array initialised to -1
// allows doom to have a "no value set yet" boolean value!
int r;
for (r = 14; r >= 0; r -= 2)
{
if (!sprtemp[frame].Texture[r].isValid())
{
sprtemp[frame].Texture[r] = lump;
if (flipped)
{
sprtemp[frame].Flip |= 1 << r;
}
sprtemp[frame].rotate = false;
}
}
}
else
{
if (rotation <= 8)
{
rotation = (rotation - 1) * 2;
}
else
{
rotation = (rotation - 9) * 2 + 1;
}
if (!sprtemp[frame].Texture[rotation].isValid())
{
// the lump is only used for one rotation
sprtemp[frame].Texture[rotation] = lump;
if (flipped)
{
sprtemp[frame].Flip |= 1 << rotation;
}
sprtemp[frame].rotate = true;
}
}
}
// [RH] Seperated out of R_InitSpriteDefs()
static void R_InstallSprite (int num)
{
int frame;
int framestart;
int rot;
// int undefinedFix;
if (maxframe == -1)
{
sprites[num].numframes = 0;
return;
}
maxframe++;
// [RH] If any frames are undefined, but there are some defined frames, map
// them to the first defined frame. This is a fix for Doom Raider, which actually
// worked with ZDoom 2.0.47, because of a bug here. It does not define frames A,
// B, or C for the sprite PSBG, but because I had sprtemp[].rotate defined as a
// bool, this code never detected that it was not actually present. After switching
// to the unified texture system, this caused it to crash while loading the wad.
// [RH] Let undefined frames actually be blank because LWM uses this in at least
// one of her wads.
// for (frame = 0; frame < maxframe && sprtemp[frame].rotate == -1; ++frame)
// { }
//
// undefinedFix = frame;
for (frame = 0; frame < maxframe; ++frame)
{
switch (sprtemp[frame].rotate)
{
case -1:
// no rotations were found for that frame at all
//I_FatalError ("R_InstallSprite: No patches found for %s frame %c", sprites[num].name, frame+'A');
break;
case 0:
// only the first rotation is needed
for (rot = 1; rot < 16; ++rot)
{
sprtemp[frame].Texture[rot] = sprtemp[frame].Texture[0];
}
// If the frame is flipped, they all should be
if (sprtemp[frame].Flip & 1)
{
sprtemp[frame].Flip = 0xFFFF;
}
break;
case 1:
// must have all 8 frame pairs
for (rot = 0; rot < 8; ++rot)
{
if (!sprtemp[frame].Texture[rot*2+1].isValid())
{
sprtemp[frame].Texture[rot*2+1] = sprtemp[frame].Texture[rot*2];
if (sprtemp[frame].Flip & (1 << (rot*2)))
{
sprtemp[frame].Flip |= 1 << (rot*2+1);
}
}
if (!sprtemp[frame].Texture[rot*2].isValid())
{
sprtemp[frame].Texture[rot*2] = sprtemp[frame].Texture[rot*2+1];
if (sprtemp[frame].Flip & (1 << (rot*2+1)))
{
sprtemp[frame].Flip |= 1 << (rot*2);
}
}
}
for (rot = 0; rot < 16; ++rot)
{
if (!sprtemp[frame].Texture[rot].isValid())
I_FatalError ("R_InstallSprite: Sprite %s frame %c is missing rotations",
sprites[num].name, frame+'A');
}
break;
}
}
for (frame = 0; frame < maxframe; ++frame)
{
if (sprtemp[frame].rotate == -1)
{
memset (&sprtemp[frame], 0, sizeof(sprtemp[0]));
}
}
// allocate space for the frames present and copy sprtemp to it
sprites[num].numframes = maxframe;
sprites[num].spriteframes = WORD(framestart = SpriteFrames.Reserve (maxframe));
for (frame = 0; frame < maxframe; ++frame)
{
memcpy (SpriteFrames[framestart+frame].Texture, sprtemp[frame].Texture, sizeof(sprtemp[frame].Texture));
SpriteFrames[framestart+frame].Flip = sprtemp[frame].Flip;
}
// Let the textures know about the rotations
for (frame = 0; frame < maxframe; ++frame)
{
if (sprtemp[frame].rotate == 1)
{
for (int rot = 0; rot < 16; ++rot)
{
TexMan[sprtemp[frame].Texture[rot]]->Rotations = framestart + frame;
}
}
}
}
//
// R_InitSpriteDefs
// Pass a null terminated list of sprite names
// (4 chars exactly) to be used.
// Builds the sprite rotation matrices to account
// for horizontally flipped sprites.
// Will report an error if the lumps are inconsistant.
// Only called at startup.
//
// Sprite lump names are 4 characters for the actor,
// a letter for the frame, and a number for the rotation.
// A sprite that is flippable will have an additional
// letter/number appended.
// The rotation character can be 0 to signify no rotations.
//
void R_InitSpriteDefs ()
{
struct Hasher
{
int Head, Next;
} *hashes;
unsigned int i, max;
DWORD intname;
// Create a hash table to speed up the process
max = TexMan.NumTextures();
hashes = (Hasher *)alloca (sizeof(Hasher) * max);
for (i = 0; i < max; ++i)
{
hashes[i].Head = -1;
}
for (i = 0; i < max; ++i)
{
FTexture *tex = TexMan.ByIndex(i);
if (tex->UseType == FTexture::TEX_Sprite && strlen(tex->Name) >= 6)
{
DWORD bucket = tex->dwName % max;
hashes[i].Next = hashes[bucket].Head;
hashes[bucket].Head = i;
}
}
// scan all the lump names for each of the names, noting the highest frame letter.
for (i = 0; i < sprites.Size(); ++i)
{
memset (sprtemp, 0xFF, sizeof(sprtemp));
for (int j = 0; j < MAX_SPRITE_FRAMES; ++j)
{
sprtemp[j].Flip = 0;
}
maxframe = -1;
intname = sprites[i].dwName;
// scan the lumps, filling in the frames for whatever is found
int hash = hashes[intname % max].Head;
while (hash != -1)
{
FTexture *tex = TexMan[hash];
if (tex->dwName == intname)
{
R_InstallSpriteLump (FTextureID(hash), tex->Name[4] - 'A', tex->Name[5], false);
if (tex->Name[6])
R_InstallSpriteLump (FTextureID(hash), tex->Name[6] - 'A', tex->Name[7], true);
}
hash = hashes[hash].Next;
}
R_InstallSprite ((int)i);
}
}
// [RH]
// R_InitSkins
// Reads in everything applicable to a skin. The skins should have already
// been counted and had their identifiers assigned to namespaces.
//
#define NUMSKINSOUNDS 17
static const char *skinsoundnames[NUMSKINSOUNDS][2] =
{ // The *painXXX sounds must be the first four
{ "dsplpain", "*pain100" },
{ "dsplpain", "*pain75" },
{ "dsplpain", "*pain50" },
{ "dsplpain", "*pain25" },
{ "dsplpain", "*poison" },
{ "dsoof", "*grunt" },
{ "dsoof", "*land" },
{ "dspldeth", "*death" },
{ "dspldeth", "*wimpydeath" },
{ "dspdiehi", "*xdeath" },
{ "dspdiehi", "*crazydeath" },
{ "dsnoway", "*usefail" },
{ "dsnoway", "*puzzfail" },
{ "dsslop", "*gibbed" },
{ "dsslop", "*splat" },
{ "dspunch", "*fist" },
{ "dsjump", "*jump" }
};
/*
static int STACK_ARGS skinsorter (const void *a, const void *b)
{
return stricmp (((FPlayerSkin *)a)->name, ((FPlayerSkin *)b)->name);
}
*/
void R_InitSkins (void)
{
FSoundID playersoundrefs[NUMSKINSOUNDS];
spritedef_t temp;
int sndlumps[NUMSKINSOUNDS];
char key[65];
DWORD intname, crouchname;
size_t i;
int j, k, base;
int lastlump;
int aliasid;
bool remove;
const PClass *basetype, *transtype;
key[sizeof(key)-1] = 0;
i = PlayerClasses.Size () - 1;
lastlump = 0;
for (j = 0; j < NUMSKINSOUNDS; ++j)
{
playersoundrefs[j] = skinsoundnames[j][1];
}
while ((base = Wads.FindLump ("S_SKIN", &lastlump, true)) != -1)
{
// The player sprite has 23 frames. This means that the S_SKIN
// marker needs a minimum of 23 lumps after it.
if (base >= Wads.GetNumLumps() - 23 || base == -1)
continue;
i++;
for (j = 0; j < NUMSKINSOUNDS; j++)
sndlumps[j] = -1;
skins[i].namespc = Wads.GetLumpNamespace (base);
FScanner sc(base);
intname = 0;
crouchname = 0;
remove = false;
basetype = NULL;
transtype = NULL;
// Data is stored as "key = data".
- 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 ())
{
- 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
strncpy (key, sc.String, sizeof(key)-1);
if (!sc.GetString() || sc.String[0] != '=')
{
Printf (PRINT_BOLD, "Bad format for skin %d: %s\n", (int)i, key);
break;
}
- 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.GetString ();
if (0 == stricmp (key, "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
strncpy (skins[i].name, sc.String, 16);
for (j = 0; (size_t)j < i; j++)
{
if (stricmp (skins[i].name, skins[j].name) == 0)
{
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 (skins[i].name, countof(skins[i].name), "skin%d", (int)i);
Printf (PRINT_BOLD, "Skin %s duplicated as %s\n",
skins[j].name, skins[i].name);
break;
}
}
}
else if (0 == stricmp (key, "sprite"))
{
for (j = 3; j >= 0; j--)
- 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.String[j] = toupper (sc.String[j]);
intname = *((DWORD *)sc.String);
}
else if (0 == stricmp (key, "crouchsprite"))
{
for (j = 3; j >= 0; j--)
- 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.String[j] = toupper (sc.String[j]);
crouchname = *((DWORD *)sc.String);
}
else if (0 == stricmp (key, "face"))
{
for (j = 2; j >= 0; j--)
- 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
skins[i].face[j] = toupper (sc.String[j]);
skins[i].face[3] = '\0';
}
else if (0 == stricmp (key, "gender"))
{
- 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
skins[i].gender = D_GenderToInt (sc.String);
}
else if (0 == stricmp (key, "scale"))
{
skins[i].ScaleX = clamp<fixed_t> (FLOAT2FIXED(atof (sc.String)), 1, 256*FRACUNIT);
skins[i].ScaleY = skins[i].ScaleX;
}
else if (0 == stricmp (key, "game"))
{
if (gameinfo.gametype == GAME_Heretic)
basetype = PClass::FindClass (NAME_HereticPlayer);
else if (gameinfo.gametype == GAME_Strife)
basetype = PClass::FindClass (NAME_StrifePlayer);
else
basetype = PClass::FindClass (NAME_DoomPlayer);
transtype = basetype;
- 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 (stricmp (sc.String, "heretic") == 0)
{
if (gameinfo.gametype & GAME_DoomChex)
{
transtype = PClass::FindClass (NAME_HereticPlayer);
skins[i].othergame = true;
}
else if (gameinfo.gametype != GAME_Heretic)
{
remove = true;
}
}
- 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 (stricmp (sc.String, "strife") == 0)
{
if (gameinfo.gametype != GAME_Strife)
{
remove = true;
}
}
else
{
if (gameinfo.gametype == GAME_Heretic)
{
transtype = PClass::FindClass (NAME_DoomPlayer);
skins[i].othergame = true;
}
else if (!(gameinfo.gametype & GAME_DoomChex))
{
remove = true;
}
}
if (remove)
break;
}
else if (0 == stricmp (key, "class"))
{ // [GRB] Define the skin for a specific player class
- 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
int pclass = D_PlayerClassToInt (sc.String);
if (pclass < 0)
{
remove = true;
break;
}
basetype = transtype = PlayerClasses[pclass].Type;
}
else if (key[0] == '*')
{ // Player sound replacment (ZDoom extension)
- 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
int lump = Wads.CheckNumForName (sc.String, skins[i].namespc);
if (lump == -1)
{
lump = Wads.CheckNumForFullName (sc.String, true, ns_sounds);
}
if (lump != -1)
{
if (stricmp (key, "*pain") == 0)
{ // Replace all pain sounds in one go
aliasid = S_AddPlayerSound (skins[i].name, skins[i].gender,
playersoundrefs[0], lump, true);
for (int l = 3; l > 0; --l)
{
S_AddPlayerSoundExisting (skins[i].name, skins[i].gender,
playersoundrefs[l], aliasid, true);
}
}
else
{
int sndref = S_FindSoundNoHash (key);
if (sndref != 0)
{
S_AddPlayerSound (skins[i].name, skins[i].gender, sndref, lump, true);
}
}
}
}
else
{
for (j = 0; j < NUMSKINSOUNDS; j++)
{
if (stricmp (key, skinsoundnames[j][0]) == 0)
{
- 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
sndlumps[j] = Wads.CheckNumForName (sc.String, skins[i].namespc);
if (sndlumps[j] == -1)
{ // Replacement not found, try finding it in the global namespace
sndlumps[j] = Wads.CheckNumForFullName (sc.String, true, ns_sounds);
}
}
}
//if (j == 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
// Printf ("Funny info for skin %i: %s = %s\n", i, key, sc.String);
}
}
// [GRB] Assume Doom skin by default
if (!remove && basetype == NULL)
{
if (gameinfo.gametype & GAME_DoomChex)
{
basetype = transtype = PClass::FindClass (NAME_DoomPlayer);
}
else if (gameinfo.gametype == GAME_Heretic)
{
basetype = PClass::FindClass (NAME_HereticPlayer);
transtype = PClass::FindClass (NAME_DoomPlayer);
skins[i].othergame = true;
}
else
{
remove = true;
}
}
if (!remove)
{
skins[i].range0start = transtype->Meta.GetMetaInt (APMETA_ColorRange) & 0xff;
skins[i].range0end = transtype->Meta.GetMetaInt (APMETA_ColorRange) >> 8;
remove = true;
for (j = 0; j < (int)PlayerClasses.Size (); j++)
{
const PClass *type = PlayerClasses[j].Type;
if (type->IsDescendantOf (basetype) &&
GetDefaultByType (type)->SpawnState->sprite == GetDefaultByType (basetype)->SpawnState->sprite &&
type->Meta.GetMetaInt (APMETA_ColorRange) == basetype->Meta.GetMetaInt (APMETA_ColorRange))
{
PlayerClasses[j].Skins.Push ((int)i);
remove = false;
}
}
}
if (!remove)
{
if (skins[i].name[0] == 0)
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 (skins[i].name, countof(skins[i].name), "skin%d", (int)i);
// Now collect the sprite frames for this skin. If the sprite name was not
// specified, use whatever immediately follows the specifier lump.
if (intname == 0)
{
char name[9];
Wads.GetLumpName (name, base+1);
memcpy(&intname, name, 4);
}
int basens = Wads.GetLumpNamespace(base);
for(int spr = 0; spr<2; spr++)
{
memset (sprtemp, 0xFFFF, sizeof(sprtemp));
for (k = 0; k < MAX_SPRITE_FRAMES; ++k)
{
sprtemp[k].Flip = 0;
}
maxframe = -1;
if (spr == 1)
{
if (crouchname !=0 && crouchname != intname)
{
intname = crouchname;
}
else
{
skins[i].crouchsprite = -1;
break;
}
}
for (k = base + 1; Wads.GetLumpNamespace(k) == basens; k++)
{
char lname[9];
DWORD lnameint;
Wads.GetLumpName (lname, k);
memcpy(&lnameint, lname, 4);
if (lnameint == intname)
{
FTextureID picnum = TexMan.CreateTexture(k, FTexture::TEX_SkinSprite);
R_InstallSpriteLump (picnum, lname[4] - 'A', lname[5], false);
if (lname[6])
R_InstallSpriteLump (picnum, lname[6] - 'A', lname[7], true);
}
}
if (spr == 0 && maxframe <= 0)
{
Printf (PRINT_BOLD, "Skin %s (#%d) has no frames. Removing.\n", skins[i].name, (int)i);
remove = true;
break;
}
Wads.GetLumpName (temp.name, base+1);
temp.name[4] = 0;
int sprno = (int)sprites.Push (temp);
if (spr==0) skins[i].sprite = sprno;
else skins[i].crouchsprite = sprno;
R_InstallSprite (sprno);
}
}
if (remove)
{
if (i < numskins-1)
memmove (&skins[i], &skins[i+1], sizeof(skins[0])*(numskins-i-1));
i--;
continue;
}
// Register any sounds this skin provides
aliasid = 0;
for (j = 0; j < NUMSKINSOUNDS; j++)
{
if (sndlumps[j] != -1)
{
if (j == 0 || sndlumps[j] != sndlumps[j-1])
{
aliasid = S_AddPlayerSound (skins[i].name, skins[i].gender,
playersoundrefs[j], sndlumps[j], true);
}
else
{
S_AddPlayerSoundExisting (skins[i].name, skins[i].gender,
playersoundrefs[j], aliasid, true);
}
}
}
// Make sure face prefix is a full 3 chars
if (skins[i].face[1] == 0 || skins[i].face[2] == 0)
{
skins[i].face[0] = 0;
}
}
if (numskins > PlayerClasses.Size ())
{ // The sound table may have changed, so rehash it.
S_HashSounds ();
S_ShrinkPlayerSoundLists ();
}
}
// [RH] Find a skin by name
int R_FindSkin (const char *name, int pclass)
{
if (stricmp ("base", name) == 0)
{
return pclass;
}
for (unsigned i = PlayerClasses.Size(); i < numskins; i++)
{
if (strnicmp (skins[i].name, name, 16) == 0)
{
if (PlayerClasses[pclass].CheckSkin (i))
return i;
else
return pclass;
}
}
return pclass;
}
// [RH] List the names of all installed skins
CCMD (skins)
{
int i;
for (i = PlayerClasses.Size ()-1; i < (int)numskins; i++)
Printf ("% 3d %s\n", i-PlayerClasses.Size ()+1, skins[i].name);
}
//
// GAME FUNCTIONS
//
int MaxVisSprites;
vissprite_t **vissprites;
vissprite_t **firstvissprite;
vissprite_t **vissprite_p;
vissprite_t **lastvissprite;
int newvissprite;
static vissprite_t **spritesorter;
static int spritesortersize = 0;
static int vsprcount;
- Yay! We now seem to be free of memory leaks! The next step will be to merge a lot of these static destructor-only structs into regular functions added to the exit chain with atterm so that they can be called in a deterministic order and not whatever order the linker decides to put them in. (Interestingly, the amount of memory used when repeatedly executing the same map command at the console varies up and down, but it now stays relatively stable rather than increasing unbounded.) - Fixed: The list of resolutions in the video modes menu was not freed at exit. - Fixed: mus_playing.name was not freed at exit. - Fixed: SN_StopAllSequences() should be called at the start of P_FreeLevelData(), not just before the call to P_SetupLevel() in G_DoLoadLevel(), so it can run even at exit. And C_FullConsole() can call P_FreeLevelData() to free more memory too. - Fixed: StatusBar was not freed at exit. - Fixed: spritesorter was not freed at exit. - Fixed: Bad things happened if FString's data pool was destroyed before all C_RemoveTabCommand() calls were made. - Added an overload for FArchive << FString. - Fixed: The players' log text was not freed at exit. - Fixed: Bot information was not freed at exit. - Fixed: doomcom was not freed at exit. But since it's always created, there's no reason why it needs to be allocated from the heap. My guess is that in the DOS days, the external packet driver was responsible for allocating doomcom and passed its location with the -net parameter. - Fixed: FBlockNodes were not freed at exit. - Fixed: Openings were not freed at exit. - Fixed: Drawsegs were not freed at exit. - Fixed: Vissprites were not freed at exit. - Fixed: Console command history was not freed at exit. - Fixed: Visplanes were not freed at exit. - Fixed: Call P_FreeLevelData() at exit. - Fixed: Channel, SoundCurve, and PlayList in s_sound.cpp were not freed at exit. - Fixed: Sound sequences were not freed at exit. - Fixed: DSeqNode::Serialize() did not resize the m_SequenceChoices array when loading. SVN r106 (trunk)
2006-05-11 04:00:58 +00:00
static void R_CreateSkinTranslation (const char *palname)
{
FMemLump lump = Wads.ReadLump (palname);
const BYTE *otherPal = (BYTE *)lump.GetMem();
for (int i = 0; i < 256; ++i)
{
OtherGameSkinRemap[i] = ColorMatcher.Pick (otherPal[0], otherPal[1], otherPal[2]);
- 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
OtherGameSkinPalette[i] = PalEntry(otherPal[0], otherPal[1], otherPal[2]);
otherPal += 3;
}
}
//
// R_InitSprites
// Called at program start.
//
void R_InitSprites ()
{
int lump, lastlump;
unsigned int i, j;
clearbufshort (zeroarray, MAXWIDTH, 0);
// [RH] Create a standard translation to map skins between Heretic and Doom
if (gameinfo.gametype == GAME_DoomChex)
{
R_CreateSkinTranslation ("SPALHTIC");
}
else
{
R_CreateSkinTranslation ("SPALDOOM");
}
// [RH] Count the number of skins.
numskins = PlayerClasses.Size ();
lastlump = 0;
while ((lump = Wads.FindLump ("S_SKIN", &lastlump, true)) != -1)
{
numskins++;
}
// [RH] Do some preliminary setup
skins = new FPlayerSkin[numskins];
memset (skins, 0, sizeof(*skins) * numskins);
for (i = 0; i < numskins; i++)
{ // Assume Doom skin by default
const PClass *type = PlayerClasses[0].Type;
skins[i].range0start = type->Meta.GetMetaInt (APMETA_ColorRange) & 255;
skins[i].range0end = type->Meta.GetMetaInt (APMETA_ColorRange) >> 8;
skins[i].ScaleX = GetDefaultByType (type)->scaleX;
skins[i].ScaleY = GetDefaultByType (type)->scaleY;
}
R_InitSpriteDefs ();
NumStdSprites = sprites.Size();
R_InitSkins (); // [RH] Finish loading skin data
// [RH] Set up base skin
// [GRB] Each player class has its own base skin
for (i = 0; i < PlayerClasses.Size (); i++)
{
const PClass *basetype = PlayerClasses[i].Type;
const char *pclassface = basetype->Meta.GetMetaString (APMETA_Face);
strcpy (skins[i].name, "Base");
if (pclassface == NULL || strcmp(pclassface, "None") == 0)
{
skins[i].face[0] = 'S';
skins[i].face[1] = 'T';
skins[i].face[2] = 'F';
skins[i].face[3] = '\0';
}
else
{
strcpy(skins[i].face, pclassface);
}
skins[i].range0start = basetype->Meta.GetMetaInt (APMETA_ColorRange) & 255;
skins[i].range0end = basetype->Meta.GetMetaInt (APMETA_ColorRange) >> 8;
skins[i].ScaleX = GetDefaultByType (basetype)->scaleX;
skins[i].ScaleY = GetDefaultByType (basetype)->scaleY;
skins[i].sprite = GetDefaultByType (basetype)->SpawnState->sprite;
skins[i].namespc = ns_global;
PlayerClasses[i].Skins.Push (i);
if (memcmp (sprites[skins[i].sprite].name, "PLAY", 4) == 0)
{
for (j = 0; j < sprites.Size (); j++)
{
if (memcmp (sprites[j].name, deh.PlayerSprite, 4) == 0)
{
skins[i].sprite = (int)j;
break;
}
}
}
}
// [RH] Sort the skins, but leave base as skin 0
//qsort (&skins[PlayerClasses.Size ()], numskins-PlayerClasses.Size (), sizeof(FPlayerSkin), skinsorter);
}
void R_DeinitSprites()
{
// Free skins
if (skins != NULL)
{
delete[] skins;
skins = NULL;
}
// Free vissprites
for (int i = 0; i < MaxVisSprites; ++i)
{
delete vissprites[i];
}
free (vissprites);
vissprites = NULL;
vissprite_p = lastvissprite = NULL;
MaxVisSprites = 0;
// Free vissprites sorter
if (spritesorter != NULL)
{
delete[] spritesorter;
spritesortersize = 0;
spritesorter = NULL;
}
}
//
// R_ClearSprites
// Called at frame start.
//
void R_ClearSprites (void)
{
vissprite_p = firstvissprite;
}
//
// R_NewVisSprite
//
vissprite_t *R_NewVisSprite (void)
{
if (vissprite_p == lastvissprite)
{
ptrdiff_t firstvisspritenum = firstvissprite - vissprites;
ptrdiff_t prevvisspritenum = vissprite_p - vissprites;
MaxVisSprites = MaxVisSprites ? MaxVisSprites * 2 : 128;
vissprites = (vissprite_t **)M_Realloc (vissprites, MaxVisSprites * sizeof(vissprite_t));
lastvissprite = &vissprites[MaxVisSprites];
firstvissprite = &vissprites[firstvisspritenum];
vissprite_p = &vissprites[prevvisspritenum];
DPrintf ("MaxVisSprites increased to %d\n", MaxVisSprites);
// Allocate sprites from the new pile
for (vissprite_t **p = vissprite_p; p < lastvissprite; ++p)
{
*p = new vissprite_t;
}
}
vissprite_p++;
return *(vissprite_p-1);
}
//
// R_DrawMaskedColumn
// Used for sprites and masked mid textures.
// Masked means: partly transparent, i.e. stored
// in posts/runs of opaque pixels.
//
short* mfloorclip;
short* mceilingclip;
fixed_t spryscale;
fixed_t sprtopscreen;
bool sprflipvert;
void R_DrawMaskedColumn (const BYTE *column, const FTexture::Span *span)
{
while (span->Length != 0)
{
const int length = span->Length;
const int top = span->TopOffset;
// calculate unclipped screen coordinates for post
dc_yl = (sprtopscreen + spryscale * top) >> FRACBITS;
dc_yh = (sprtopscreen + spryscale * (top + length) - FRACUNIT) >> FRACBITS;
if (sprflipvert)
{
swap (dc_yl, dc_yh);
}
if (dc_yh >= mfloorclip[dc_x])
{
dc_yh = mfloorclip[dc_x] - 1;
}
if (dc_yl < mceilingclip[dc_x])
{
dc_yl = mceilingclip[dc_x];
}
if (dc_yl <= dc_yh)
{
if (sprflipvert)
{
dc_texturefrac = (dc_yl*dc_iscale) - (top << FRACBITS)
- FixedMul (centeryfrac, dc_iscale) - dc_texturemid;
const fixed_t maxfrac = length << FRACBITS;
while (dc_texturefrac >= maxfrac)
{
if (++dc_yl > dc_yh)
goto nextpost;
dc_texturefrac += dc_iscale;
}
fixed_t endfrac = dc_texturefrac + (dc_yh-dc_yl)*dc_iscale;
while (endfrac < 0)
{
if (--dc_yh < dc_yl)
goto nextpost;
endfrac -= dc_iscale;
}
}
else
{
dc_texturefrac = dc_texturemid - (top << FRACBITS)
+ (dc_yl*dc_iscale) - FixedMul (centeryfrac-FRACUNIT, dc_iscale);
while (dc_texturefrac < 0)
{
if (++dc_yl > dc_yh)
goto nextpost;
dc_texturefrac += dc_iscale;
}
fixed_t endfrac = dc_texturefrac + (dc_yh-dc_yl)*dc_iscale;
const fixed_t maxfrac = length << FRACBITS;
if (dc_yh < mfloorclip[dc_x]-1 && endfrac < maxfrac - dc_iscale)
{
dc_yh++;
}
else while (endfrac >= maxfrac)
{
if (--dc_yh < dc_yl)
goto nextpost;
endfrac -= dc_iscale;
}
}
dc_source = column + top;
dc_dest = ylookup[dc_yl] + dc_x + dc_destorg;
dc_count = dc_yh - dc_yl + 1;
colfunc ();
}
nextpost:
span++;
}
}
//
// R_DrawVisSprite
// mfloorclip and mceilingclip should also be set.
//
void R_DrawVisSprite (vissprite_t *vis)
{
const BYTE *pixels;
const FTexture::Span *spans;
fixed_t frac;
FTexture *tex;
int x2, stop4;
fixed_t xiscale;
ESPSResult mode;
dc_colormap = vis->colormap;
- 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
mode = R_SetPatchStyle (vis->RenderStyle, vis->alpha, vis->Translation, vis->FillColor);
if (mode != DontDraw)
{
if (mode == DoDraw0)
{
// One column at a time
stop4 = vis->x1;
}
else // DoDraw1
{
// Up to four columns at a time
stop4 = (vis->x2 + 1) & ~3;
}
tex = vis->pic;
spryscale = vis->yscale;
sprflipvert = false;
dc_iscale = 0xffffffffu / (unsigned)vis->yscale;
dc_texturemid = vis->texturemid;
frac = vis->startfrac;
xiscale = vis->xiscale;
sprtopscreen = centeryfrac - FixedMul (dc_texturemid, spryscale);
dc_x = vis->x1;
x2 = vis->x2 + 1;
if (dc_x < x2)
{
while ((dc_x < stop4) && (dc_x & 3))
{
pixels = tex->GetColumn (frac >> FRACBITS, &spans);
R_DrawMaskedColumn (pixels, spans);
dc_x++;
frac += xiscale;
}
while (dc_x < stop4)
{
rt_initcols();
for (int zz = 4; zz; --zz)
{
pixels = tex->GetColumn (frac >> FRACBITS, &spans);
R_DrawMaskedColumnHoriz (pixels, spans);
dc_x++;
frac += xiscale;
}
rt_draw4cols (dc_x - 4);
}
while (dc_x < x2)
{
pixels = tex->GetColumn (frac >> FRACBITS, &spans);
R_DrawMaskedColumn (pixels, spans);
dc_x++;
frac += xiscale;
}
}
}
R_FinishSetPatchStyle ();
NetUpdate ();
}
//
// R_ProjectSprite
// Generates a vissprite for a thing if it might be visible.
//
void R_ProjectSprite (AActor *thing, int fakeside)
{
fixed_t fx, fy, fz;
fixed_t tr_x;
fixed_t tr_y;
fixed_t gzt; // killough 3/27/98
fixed_t gzb; // [RH] use bottom of sprite, not actor
fixed_t tx, tx2;
fixed_t tz;
fixed_t xscale;
int x1;
int x2;
FTextureID picnum;
FTexture *tex;
WORD flip;
vissprite_t* vis;
fixed_t iscale;
sector_t* heightsec; // killough 3/27/98
- 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
// Don't waste time projecting sprites that are definitely not visible.
if (thing == NULL ||
(thing->renderflags & RF_INVISIBLE) ||
- 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
!thing->RenderStyle.IsVisible(thing->alpha))
{
return;
}
// [RH] Interpolate the sprite's position to make it look smooth
fx = thing->PrevX + FixedMul (r_TicFrac, thing->x - thing->PrevX);
fy = thing->PrevY + FixedMul (r_TicFrac, thing->y - thing->PrevY);
fz = thing->PrevZ + FixedMul (r_TicFrac, thing->z - thing->PrevZ);
// transform the origin point
tr_x = fx - viewx;
tr_y = fy - viewy;
tz = DMulScale20 (tr_x, viewtancos, tr_y, viewtansin);
// thing is behind view plane?
if (tz < MINZ)
return;
tx = DMulScale16 (tr_x, viewsin, -tr_y, viewcos);
// [RH] Flip for mirrors
if (MirrorFlags & RF_XFLIP)
{
tx = -tx;
}
tx2 = tx >> 4;
// too far off the side?
if ((abs (tx) >> 6) > tz)
{
return;
}
xscale = DivScale12 (centerxfrac, tz);
if (thing->picnum.isValid())
{
picnum = thing->picnum;
tex = TexMan(picnum);
if (tex->UseType == FTexture::TEX_Null)
{
return;
}
flip = 0;
if (tex->Rotations != 0xFFFF)
{
// choose a different rotation based on player view
spriteframe_t *sprframe = &SpriteFrames[tex->Rotations];
angle_t ang = R_PointToAngle (fx, fy);
angle_t rot;
if (sprframe->Texture[0] == sprframe->Texture[1])
{
rot = (ang - thing->angle + (angle_t)(ANGLE_45/2)*9) >> 28;
}
else
{
rot = (ang - thing->angle + (angle_t)(ANGLE_45/2)*9-(angle_t)(ANGLE_180/16)) >> 28;
}
picnum = sprframe->Texture[rot];
flip = sprframe->Flip & (1 << rot);
tex = TexMan[picnum]; // Do not animate the rotation
}
}
else
{
// decide which texture to use for the sprite
#ifdef RANGECHECK
if ((unsigned)thing->sprite >= (unsigned)sprites.Size ())
{
DPrintf ("R_ProjectSprite: invalid sprite number %i\n", thing->sprite);
return;
}
#endif
spritedef_t *sprdef = &sprites[thing->sprite];
if (thing->frame >= sprdef->numframes)
{
// If there are no frames at all for this sprite, don't draw it.
return;
}
else
{
//picnum = SpriteFrames[sprdef->spriteframes + thing->frame].Texture[0];
// choose a different rotation based on player view
spriteframe_t *sprframe = &SpriteFrames[sprdef->spriteframes + thing->frame];
angle_t ang = R_PointToAngle (fx, fy);
angle_t rot;
if (sprframe->Texture[0] == sprframe->Texture[1])
{
rot = (ang - thing->angle + (angle_t)(ANGLE_45/2)*9) >> 28;
}
else
{
rot = (ang - thing->angle + (angle_t)(ANGLE_45/2)*9-(angle_t)(ANGLE_180/16)) >> 28;
}
picnum = sprframe->Texture[rot];
flip = sprframe->Flip & (1 << rot);
tex = TexMan[picnum]; // Do not animate the rotation
}
}
if (tex == NULL || tex->UseType == FTexture::TEX_Null)
{
return;
}
// [RH] Added scaling
int scaled_to = tex->GetScaledTopOffset();
int scaled_bo = scaled_to - tex->GetScaledHeight();
gzt = fz + thing->scaleY * scaled_to;
gzb = fz + thing->scaleY * scaled_bo;
// [RH] Reject sprites that are off the top or bottom of the screen
if (MulScale12 (globaluclip, tz) > viewz - gzb ||
MulScale12 (globaldclip, tz) < viewz - gzt)
{
return;
}
// [RH] Flip for mirrors and renderflags
if ((MirrorFlags ^ thing->renderflags) & RF_XFLIP)
{
flip = !flip;
}
// calculate edges of the shape
const fixed_t thingxscalemul = DivScale16(thing->scaleX, tex->xScale);
tx -= (flip ? (tex->GetWidth() - tex->LeftOffset - 1) : tex->LeftOffset) * thingxscalemul;
x1 = centerx + MulScale32 (tx, xscale);
// off the right side?
if (x1 > WindowRight)
return;
tx += tex->GetWidth() * thingxscalemul;
x2 = centerx + MulScale32 (tx, xscale);
// off the left side or too small?
if (x2 < WindowLeft || x2 <= x1)
return;
xscale = FixedDiv(FixedMul(thing->scaleX, xscale), tex->xScale);
iscale = (tex->GetWidth() << FRACBITS) / (x2 - x1);
x2--;
// killough 3/27/98: exclude things totally separated
// from the viewer, by either water or fake ceilings
// killough 4/11/98: improve sprite clipping for underwater/fake ceilings
heightsec = thing->Sector->GetHeightSec();
if (heightsec != NULL) // only clip things which are in special sectors
{
if (fakeside == FAKED_AboveCeiling)
{
if (gzt < heightsec->ceilingplane.ZatPoint (fx, fy))
return;
}
else if (fakeside == FAKED_BelowFloor)
{
if (gzb >= heightsec->floorplane.ZatPoint (fx, fy))
return;
}
else
{
if (gzt < heightsec->floorplane.ZatPoint (fx, fy))
return;
if (gzb >= heightsec->ceilingplane.ZatPoint (fx, fy))
return;
}
}
// store information in a vissprite
vis = R_NewVisSprite ();
// killough 3/27/98: save sector for special clipping later
vis->heightsec = heightsec;
vis->sector = thing->Sector;
fixed_t yscale = DivScale16(thing->scaleY, tex->yScale);
vis->renderflags = thing->renderflags;
vis->RenderStyle = thing->RenderStyle;
- 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
vis->FillColor = thing->fillcolor;
vis->xscale = xscale;
vis->yscale = Scale (InvZtoScale, yscale, tz << 4);
vis->depth = tz;
vis->idepth = (DWORD)DivScale32 (1, tz) >> 1; // tz is 20.12, so idepth ought to be 12.20, but
vis->cx = tx2; // signed math makes it 13.19
vis->gx = fx;
vis->gy = fy;
vis->gz = gzb; // [RH] use gzb, not thing->z
vis->gzt = gzt; // killough 3/27/98
vis->floorclip = FixedDiv (thing->floorclip, yscale);
vis->texturemid = (tex->TopOffset << FRACBITS) -
FixedDiv (viewz-fz+thing->floorclip, yscale);
vis->x1 = x1 < WindowLeft ? WindowLeft : x1;
vis->x2 = x2 > WindowRight ? WindowRight : x2;
vis->Translation = thing->Translation; // [RH] thing translation table
vis->FakeFlatStat = fakeside;
vis->alpha = thing->alpha;
vis->pic = tex;
if (flip)
{
vis->startfrac = (tex->GetWidth() << FRACBITS) - 1;
vis->xiscale = -iscale;
}
else
{
vis->startfrac = 0;
vis->xiscale = iscale;
}
if (vis->x1 > x1)
vis->startfrac += vis->xiscale*(vis->x1-x1);
- 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
// The software renderer cannot invert the source without inverting the overlay
// too. That means if the source is inverted, we need to do the reverse of what
// the invert overlay flag says to do.
INTBOOL invertcolormap = (vis->RenderStyle.Flags & STYLEF_InvertOverlay);
if (vis->RenderStyle.Flags & STYLEF_InvertSource)
{
invertcolormap = !invertcolormap;
}
FDynamicColormap *mybasecolormap = basecolormap;
// Sprites that are added to the scene must fade to black.
if (vis->RenderStyle == LegacyRenderStyles[STYLE_Add] && mybasecolormap->Fade != 0)
{
mybasecolormap = GetSpecialLights(mybasecolormap->Color, 0, mybasecolormap->Desaturate);
}
- 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 (vis->RenderStyle.Flags & STYLEF_FadeToBlack)
{
if (invertcolormap)
{ // Fade to white
- 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
mybasecolormap = GetSpecialLights(mybasecolormap->Color, MAKERGB(255,255,255), mybasecolormap->Desaturate);
invertcolormap = false;
}
else
{ // Fade to black
- 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
mybasecolormap = GetSpecialLights(mybasecolormap->Color, MAKERGB(0,0,0), mybasecolormap->Desaturate);
}
}
// get light level
if (fixedcolormap != NULL)
{ // fixed map
vis->colormap = fixedcolormap;
}
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
if (invertcolormap)
{
mybasecolormap = GetSpecialLights(mybasecolormap->Color, mybasecolormap->Fade.InverseColor(), mybasecolormap->Desaturate);
}
if (fixedlightlev >= 0)
- 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
{
vis->colormap = mybasecolormap->Maps + fixedlightlev;
}
else if (!foggy && ((thing->renderflags & RF_FULLBRIGHT) || (thing->flags5 & MF5_BRIGHT)))
{ // full bright
vis->colormap = mybasecolormap->Maps;
}
else
{ // diminished light
vis->colormap = mybasecolormap->Maps + (GETPALOOKUP (
(fixed_t)DivScale12 (r_SpriteVisibility, tz), spriteshade) << COLORMAPSHIFT);
- 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
}
}
}
//
// R_AddSprites
// During BSP traversal, this adds sprites by sector.
//
// killough 9/18/98: add lightlevel as parameter, fixing underwater lighting
// [RH] Save which side of heightsec sprite is on here.
void R_AddSprites (sector_t *sec, int lightlevel, int fakeside)
{
AActor *thing;
// BSP is traversed by subsector.
// A sector might have been split into several
// subsectors during BSP building.
// Thus we check whether it was already added.
if (sec->thinglist == NULL || sec->validcount == validcount)
return;
// Well, now it will be done.
sec->validcount = validcount;
spriteshade = LIGHT2SHADE(lightlevel + r_actualextralight);
// Handle all things in sector.
for (thing = sec->thinglist; thing; thing = thing->snext)
{
R_ProjectSprite (thing, fakeside);
}
}
//
// R_DrawPSprite
//
void R_DrawPSprite (pspdef_t* psp, int pspnum, AActor *owner, fixed_t sx, fixed_t sy)
{
fixed_t tx;
int x1;
int x2;
spritedef_t* sprdef;
spriteframe_t* sprframe;
FTextureID picnum;
WORD flip;
FTexture* tex;
vissprite_t* vis;
static vissprite_t avis[NUMPSPRITES];
bool noaccel;
assert(pspnum >= 0 && pspnum < NUMPSPRITES);
// decide which patch to use
if ( (unsigned)psp->state->sprite >= (unsigned)sprites.Size ())
{
DPrintf ("R_DrawPSprite: invalid sprite number %i\n", psp->state->sprite);
return;
}
sprdef = &sprites[psp->state->sprite];
if (psp->state->GetFrame() >= sprdef->numframes)
{
DPrintf ("R_DrawPSprite: invalid sprite frame %i : %i\n", psp->state->sprite, psp->state->GetFrame());
return;
}
sprframe = &SpriteFrames[sprdef->spriteframes + psp->state->GetFrame()];
picnum = sprframe->Texture[0];
flip = sprframe->Flip & 1;
tex = TexMan(picnum);
if (tex->UseType == FTexture::TEX_Null)
return;
// calculate edges of the shape
tx = sx-((320/2)<<FRACBITS);
tx -= tex->GetScaledLeftOffset() << FRACBITS;
x1 = (centerxfrac + FixedMul (tx, pspritexscale)) >>FRACBITS;
VisPSpritesX1[pspnum] = x1;
// off the right side
if (x1 > viewwidth)
return;
tx += tex->GetScaledWidth() << FRACBITS;
x2 = ((centerxfrac + FixedMul (tx, pspritexscale)) >>FRACBITS) - 1;
// off the left side
if (x2 < 0)
return;
// store information in a vissprite
vis = &avis[pspnum];
vis->renderflags = owner->renderflags;
vis->floorclip = 0;
vis->texturemid = MulScale16((BASEYCENTER<<FRACBITS) - sy, tex->yScale) + (tex->TopOffset << FRACBITS);
if (camera->player && (RenderTarget != screen ||
- Ported vlinetallasm4 to AMD64 assembly. Even with the increased number of registers AMD64 provides, this routine still needs to be written as self- modifying code for maximum performance. The additional registers do allow for further optimization over the x86 version by allowing all four pixels to be in flight at the same time. The end result is that AMD64 ASM is about 2.18 times faster than AMD64 C and about 1.06 times faster than x86 ASM. (For further comparison, AMD64 C and x86 C are practically the same for this function.) Should I port any more assembly to AMD64, mvlineasm4 is the most likely candidate, but it's not used enough at this point to bother. Also, this may or may not work with Linux at the moment, since it doesn't have the eh_handler metadata. Win64 is easier, since I just need to structure the function prologue and epilogue properly and use some assembler directives/macros to automatically generate the metadata. And that brings up another point: You need YASM to assemble the AMD64 code, because NASM doesn't support the Win64 metadata directives. - Added an SSE version of DoBlending. This is strictly C intrinsics. VC++ still throws around unneccessary register moves. GCC seems to be pretty close to optimal, requiring only about 2 cycles/color. They're both faster than my hand-written MMX routine, so I don't need to feel bad about not hand-optimizing this for x64 builds. - Removed an extra instruction from DoBlending_MMX, transposed two instructions, and unrolled it once, shaving off about 80 cycles from the time required to blend 256 palette entries. Why? Because I tried writing a C version of the routine using compiler intrinsics and was appalled by all the extra movq's VC++ added to the code. GCC was better, but still generated extra instructions. I only wanted a C version because I can't use inline assembly with VC++'s x64 compiler, and x64 assembly is a bit of a pain. (It's a pain because Linux and Windows have different calling conventions, and you need to maintain extra metadata for functions.) So, the assembly version stays and the C version stays out. - Removed all the pixel doubling r_detail modes, since the one platform they were intended to assist (486) actually sees very little benefit from them. - Rewrote CheckMMX in C and renamed it to CheckCPU. - Fixed: CPUID function 0x80000005 is specified to return detailed L1 cache only for AMD processors, so we must not use it on other architectures, or we end up overwriting the L1 cache line size with 0 or some other number we don't actually understand. SVN r1134 (trunk)
2008-08-09 03:13:43 +00:00
viewheight == RenderTarget->GetHeight() ||
(RenderTarget->GetWidth() > 320 && !st_scale)))
{ // Adjust PSprite for fullscreen views
AWeapon *weapon = NULL;
if (camera->player != NULL)
{
weapon = camera->player->ReadyWeapon;
}
if (pspnum <= ps_flash && weapon != NULL && weapon->YAdjust != 0)
{
- Ported vlinetallasm4 to AMD64 assembly. Even with the increased number of registers AMD64 provides, this routine still needs to be written as self- modifying code for maximum performance. The additional registers do allow for further optimization over the x86 version by allowing all four pixels to be in flight at the same time. The end result is that AMD64 ASM is about 2.18 times faster than AMD64 C and about 1.06 times faster than x86 ASM. (For further comparison, AMD64 C and x86 C are practically the same for this function.) Should I port any more assembly to AMD64, mvlineasm4 is the most likely candidate, but it's not used enough at this point to bother. Also, this may or may not work with Linux at the moment, since it doesn't have the eh_handler metadata. Win64 is easier, since I just need to structure the function prologue and epilogue properly and use some assembler directives/macros to automatically generate the metadata. And that brings up another point: You need YASM to assemble the AMD64 code, because NASM doesn't support the Win64 metadata directives. - Added an SSE version of DoBlending. This is strictly C intrinsics. VC++ still throws around unneccessary register moves. GCC seems to be pretty close to optimal, requiring only about 2 cycles/color. They're both faster than my hand-written MMX routine, so I don't need to feel bad about not hand-optimizing this for x64 builds. - Removed an extra instruction from DoBlending_MMX, transposed two instructions, and unrolled it once, shaving off about 80 cycles from the time required to blend 256 palette entries. Why? Because I tried writing a C version of the routine using compiler intrinsics and was appalled by all the extra movq's VC++ added to the code. GCC was better, but still generated extra instructions. I only wanted a C version because I can't use inline assembly with VC++'s x64 compiler, and x64 assembly is a bit of a pain. (It's a pain because Linux and Windows have different calling conventions, and you need to maintain extra metadata for functions.) So, the assembly version stays and the C version stays out. - Removed all the pixel doubling r_detail modes, since the one platform they were intended to assist (486) actually sees very little benefit from them. - Rewrote CheckMMX in C and renamed it to CheckCPU. - Fixed: CPUID function 0x80000005 is specified to return detailed L1 cache only for AMD processors, so we must not use it on other architectures, or we end up overwriting the L1 cache line size with 0 or some other number we don't actually understand. SVN r1134 (trunk)
2008-08-09 03:13:43 +00:00
if (RenderTarget != screen || viewheight == RenderTarget->GetHeight())
{
vis->texturemid -= weapon->YAdjust;
}
else
{
vis->texturemid -= FixedMul (StatusBar->GetDisplacement (),
weapon->YAdjust);
}
}
}
if (pspnum <= ps_flash)
{ // Move the weapon down for 1280x1024.
vis->texturemid -= BaseRatioSizes[WidescreenRatio][2];
}
vis->x1 = x1 < 0 ? 0 : x1;
vis->x2 = x2 >= viewwidth ? viewwidth-1 : x2;
vis->xscale = DivScale16(pspritexscale, tex->xScale);
vis->yscale = DivScale16(pspriteyscale, tex->yScale);
vis->Translation = 0; // [RH] Use default colors
vis->pic = tex;
if (flip)
{
vis->xiscale = -MulScale16(pspritexiscale, tex->xScale);
vis->startfrac = (tex->GetWidth() << FRACBITS) - 1;
}
else
{
vis->xiscale = MulScale16(pspritexiscale, tex->xScale);
vis->startfrac = 0;
}
if (vis->x1 > x1)
vis->startfrac += vis->xiscale*(vis->x1-x1);
noaccel = false;
if (pspnum <= ps_flash)
{
vis->alpha = owner->alpha;
vis->RenderStyle = owner->RenderStyle;
- 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
// The software renderer cannot invert the source without inverting the overlay
// too. That means if the source is inverted, we need to do the reverse of what
// the invert overlay flag says to do.
INTBOOL invertcolormap = (vis->RenderStyle.Flags & STYLEF_InvertOverlay);
if (vis->RenderStyle.Flags & STYLEF_InvertSource)
{
invertcolormap = !invertcolormap;
}
FDynamicColormap *mybasecolormap = basecolormap;
if (vis->RenderStyle.Flags & STYLEF_FadeToBlack)
{
if (invertcolormap)
{ // Fade to white
- 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
mybasecolormap = GetSpecialLights(mybasecolormap->Color, MAKERGB(255,255,255), mybasecolormap->Desaturate);
invertcolormap = false;
}
else
{ // Fade to black
- 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
mybasecolormap = GetSpecialLights(mybasecolormap->Color, MAKERGB(0,0,0), mybasecolormap->Desaturate);
}
}
if (realfixedcolormap != NULL)
{ // fixed color
vis->colormap = realfixedcolormap->Colormap;
}
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
if (invertcolormap)
{
mybasecolormap = GetSpecialLights(mybasecolormap->Color, mybasecolormap->Fade.InverseColor(), mybasecolormap->Desaturate);
}
if (fixedlightlev >= 0)
- 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
{
vis->colormap = mybasecolormap->Maps + fixedlightlev;
}
else if (!foggy && psp->state->GetFullbright())
{ // full bright
vis->colormap = mybasecolormap->Maps; // [RH] use basecolormap
}
else
{ // local light
vis->colormap = mybasecolormap->Maps + (GETPALOOKUP (0, spriteshade) << COLORMAPSHIFT);
- 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 (camera->Inventory != NULL)
{
lighttable_t *oldcolormap = vis->colormap;
camera->Inventory->AlterWeaponSprite (vis);
if (vis->colormap != oldcolormap)
{
// The colormap has changed. Is it one we can easily identify?
// If not, then don't bother trying to identify it for
// hardware accelerated drawing.
if (vis->colormap < SpecialColormaps[0].Colormap ||
vis->colormap >= SpecialColormaps[SpecialColormaps.Size()].Colormap)
{
noaccel = true;
}
// Has the basecolormap changed? If so, we can't hardware accelerate it,
// since we don't know what it is anymore.
else if (vis->colormap < mybasecolormap->Maps ||
vis->colormap >= mybasecolormap->Maps + NUMCOLORMAPS*256)
{
noaccel = true;
}
}
}
VisPSpritesBaseColormap[pspnum] = mybasecolormap;
}
else
{
VisPSpritesBaseColormap[pspnum] = basecolormap;
vis->colormap = basecolormap->Maps;
vis->RenderStyle = STYLE_Normal;
}
// Check for hardware-assisted 2D. If it's available, and this sprite is not
// fuzzy, don't draw it until after the switch to 2D mode.
if (!noaccel && RenderTarget == screen && (DFrameBuffer *)screen->Accel2D)
{
FRenderStyle style = vis->RenderStyle;
style.CheckFuzz();
if (style.BlendOp != STYLEOP_Fuzz)
{
VisPSprites[pspnum] = vis;
return;
}
}
R_DrawVisSprite (vis);
}
//==========================================================================
//
// R_DrawPlayerSprites
//
//==========================================================================
void R_DrawPlayerSprites (void)
{
int i;
int lightnum;
pspdef_t* psp;
sector_t* sec;
static sector_t tempsec;
int floorlight, ceilinglight;
if (!r_drawplayersprites ||
!camera->player ||
(players[consoleplayer].cheats & CF_CHASECAM))
return;
2006-04-11 08:36:23 +00:00
// This used to use camera->Sector but due to interpolation that can be incorrect
// when the interpolated viewpoint is in a different sector than the camera.
sec = R_FakeFlat (viewsector, &tempsec, &floorlight,
&ceilinglight, false);
// [RH] set foggy flag
foggy = (level.fadeto || sec->ColorMap->Fade || (level.flags & LEVEL_HASFADETABLE));
r_actualextralight = foggy ? 0 : extralight << 4;
// [RH] set basecolormap
- 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
basecolormap = sec->ColorMap;
// get light level
lightnum = ((floorlight + ceilinglight) >> 1) + r_actualextralight;
spriteshade = LIGHT2SHADE(lightnum) - 24*FRACUNIT;
// clip to screen bounds
mfloorclip = screenheightarray;
mceilingclip = zeroarray;
if (camera->player != NULL)
{
fixed_t centerhack = centeryfrac;
fixed_t ofsx, ofsy;
centery = viewheight >> 1;
centeryfrac = centery << FRACBITS;
P_BobWeapon (camera->player, &camera->player->psprites[ps_weapon], &ofsx, &ofsy);
// add all active psprites
for (i = 0, psp = camera->player->psprites;
i < NUMPSPRITES;
i++, psp++)
{
// [RH] Don't draw the targeter's crosshair if the player already has a crosshair set.
if (psp->state && (i != ps_targetcenter || CrosshairImage == NULL))
{
R_DrawPSprite (psp, i, camera, psp->sx + ofsx, psp->sy + ofsy);
}
// [RH] Don't bob the targeter.
if (i == ps_flash)
{
ofsx = ofsy = 0;
}
}
centeryfrac = centerhack;
centery = centerhack >> FRACBITS;
}
}
//==========================================================================
//
// R_DrawRemainingPlayerSprites
//
// Called from D_Display to draw sprites that were not drawn by
// R_DrawPlayerSprites().
//
//==========================================================================
void R_DrawRemainingPlayerSprites()
{
for (int i = 0; i < NUMPSPRITES; ++i)
{
vissprite_t *vis;
vis = VisPSprites[i];
VisPSprites[i] = NULL;
if (vis != NULL)
{
FDynamicColormap *colormap = VisPSpritesBaseColormap[i];
bool flip = vis->xiscale < 0;
FSpecialColormap *special = NULL;
PalEntry overlay = 0;
FColormapStyle colormapstyle;
bool usecolormapstyle = false;
if (vis->colormap >= SpecialColormaps[0].Colormap &&
vis->colormap < SpecialColormaps[SpecialColormaps.Size()].Colormap)
{
// Yuck! There needs to be a better way to store colormaps in the vissprite... :(
ptrdiff_t specialmap = (vis->colormap - SpecialColormaps[0].Colormap) / sizeof(FSpecialColormap);
special = &SpecialColormaps[specialmap];
}
else if (colormap->Color == PalEntry(255,255,255) &&
colormap->Desaturate == 0)
{
overlay = colormap->Fade;
overlay.a = BYTE(((vis->colormap - colormap->Maps) >> 8) * 255 / NUMCOLORMAPS);
}
else
{
usecolormapstyle = true;
colormapstyle.Color = colormap->Color;
colormapstyle.Fade = colormap->Fade;
colormapstyle.Desaturate = colormap->Desaturate;
colormapstyle.FadeLevel = ((vis->colormap - colormap->Maps) >> 8) / float(NUMCOLORMAPS);
}
screen->DrawTexture(vis->pic,
viewwindowx + VisPSpritesX1[i],
viewwindowy + viewheight/2 - (vis->texturemid / 65536.0) * (vis->yscale / 65536.0) - 0.5,
DTA_DestWidthF, FIXED2FLOAT(vis->pic->GetWidth() * vis->xscale),
DTA_DestHeightF, FIXED2FLOAT(vis->pic->GetHeight() * vis->yscale),
DTA_Translation, TranslationToTable(vis->Translation),
DTA_FlipX, flip,
DTA_TopOffset, 0,
DTA_LeftOffset, 0,
DTA_ClipLeft, viewwindowx,
DTA_ClipTop, viewwindowy,
DTA_ClipRight, viewwindowx + viewwidth,
DTA_ClipBottom, viewwindowy + viewheight,
DTA_Alpha, vis->alpha,
DTA_RenderStyle, vis->RenderStyle,
DTA_FillColor, vis->FillColor,
DTA_SpecialColormap, special,
DTA_ColorOverlay, overlay.d,
DTA_ColormapStyle, usecolormapstyle ? &colormapstyle : NULL,
TAG_DONE);
}
}
}
//
// R_SortVisSprites
//
// [RH] The old code for this function used a bubble sort, which was far less
// than optimal with large numbers of sprites. I changed it to use the
// stdlib qsort() function instead, and now it is a *lot* faster; the
// more vissprites that need to be sorted, the better the performance
// gain compared to the old function.
//
// Sort vissprites by depth, far to near
static bool sv_compare(vissprite_t *a, vissprite_t *b)
{
return a->idepth > b->idepth;
}
#if 0
- Yay! We now seem to be free of memory leaks! The next step will be to merge a lot of these static destructor-only structs into regular functions added to the exit chain with atterm so that they can be called in a deterministic order and not whatever order the linker decides to put them in. (Interestingly, the amount of memory used when repeatedly executing the same map command at the console varies up and down, but it now stays relatively stable rather than increasing unbounded.) - Fixed: The list of resolutions in the video modes menu was not freed at exit. - Fixed: mus_playing.name was not freed at exit. - Fixed: SN_StopAllSequences() should be called at the start of P_FreeLevelData(), not just before the call to P_SetupLevel() in G_DoLoadLevel(), so it can run even at exit. And C_FullConsole() can call P_FreeLevelData() to free more memory too. - Fixed: StatusBar was not freed at exit. - Fixed: spritesorter was not freed at exit. - Fixed: Bad things happened if FString's data pool was destroyed before all C_RemoveTabCommand() calls were made. - Added an overload for FArchive << FString. - Fixed: The players' log text was not freed at exit. - Fixed: Bot information was not freed at exit. - Fixed: doomcom was not freed at exit. But since it's always created, there's no reason why it needs to be allocated from the heap. My guess is that in the DOS days, the external packet driver was responsible for allocating doomcom and passed its location with the -net parameter. - Fixed: FBlockNodes were not freed at exit. - Fixed: Openings were not freed at exit. - Fixed: Drawsegs were not freed at exit. - Fixed: Vissprites were not freed at exit. - Fixed: Console command history was not freed at exit. - Fixed: Visplanes were not freed at exit. - Fixed: Call P_FreeLevelData() at exit. - Fixed: Channel, SoundCurve, and PlayList in s_sound.cpp were not freed at exit. - Fixed: Sound sequences were not freed at exit. - Fixed: DSeqNode::Serialize() did not resize the m_SequenceChoices array when loading. SVN r106 (trunk)
2006-05-11 04:00:58 +00:00
static drawseg_t **drawsegsorter;
static int drawsegsortersize = 0;
// Sort vissprites by leftmost column, left to right
static int STACK_ARGS sv_comparex (const void *arg1, const void *arg2)
{
return (*(vissprite_t **)arg2)->x1 - (*(vissprite_t **)arg1)->x1;
}
// Sort drawsegs by rightmost column, left to right
static int STACK_ARGS sd_comparex (const void *arg1, const void *arg2)
{
return (*(drawseg_t **)arg2)->x2 - (*(drawseg_t **)arg1)->x2;
}
CVAR (Bool, r_splitsprites, true, CVAR_ARCHIVE)
// Split up vissprites that intersect drawsegs
void R_SplitVisSprites ()
{
size_t start, stop;
size_t numdrawsegs = ds_p - firstdrawseg;
size_t numsprites;
size_t spr, dseg, dseg2;
if (!r_splitsprites)
return;
if (numdrawsegs == 0 || vissprite_p - firstvissprite == 0)
return;
// Sort drawsegs from left to right
if (numdrawsegs > drawsegsortersize)
{
if (drawsegsorter != NULL)
delete[] drawsegsorter;
drawsegsortersize = numdrawsegs * 2;
drawsegsorter = new drawseg_t *[drawsegsortersize];
}
for (dseg = dseg2 = 0; dseg < numdrawsegs; ++dseg)
{
// Drawsegs that don't clip any sprites don't need to be considered.
if (firstdrawseg[dseg].silhouette)
{
drawsegsorter[dseg2++] = &firstdrawseg[dseg];
}
}
numdrawsegs = dseg2;
if (numdrawsegs == 0)
{
return;
}
qsort (drawsegsorter, numdrawsegs, sizeof(drawseg_t *), sd_comparex);
// Now sort vissprites from left to right, and walk them simultaneously
// with the drawsegs, splitting any that intersect.
start = firstvissprite - vissprites;
int p = 0;
do
{
p++;
R_SortVisSprites (sv_comparex, start);
stop = vissprite_p - vissprites;
numsprites = stop - start;
spr = dseg = 0;
do
{
vissprite_t *vis = spritesorter[spr], *vis2;
// Skip drawsegs until we get to one that doesn't end before the sprite
// begins.
while (dseg < numdrawsegs && drawsegsorter[dseg]->x2 <= vis->x1)
{
dseg++;
}
// Now split the sprite against any drawsegs it intersects
for (dseg2 = dseg; dseg2 < numdrawsegs; dseg2++)
{
drawseg_t *ds = drawsegsorter[dseg2];
if (ds->x1 > vis->x2 || ds->x2 < vis->x1)
continue;
if ((vis->idepth < ds->siz1) != (vis->idepth < ds->siz2))
{ // The drawseg is crossed; find the x where the intersection occurs
int cross = Scale (vis->idepth - ds->siz1, ds->sx2 - ds->sx1, ds->siz2 - ds->siz1) + ds->sx1 + 1;
/* if (cross < ds->x1 || cross > ds->x2)
{ // The original seg is crossed, but the drawseg is not
continue;
}
*/ if (cross <= vis->x1 || cross >= vis->x2)
{ // Don't create 0-sized sprites
continue;
}
vis->bSplitSprite = true;
// Create a new vissprite for the right part of the sprite
vis2 = R_NewVisSprite ();
*vis2 = *vis;
vis2->startfrac += vis2->xiscale * (cross - vis2->x1);
vis->x2 = cross-1;
vis2->x1 = cross;
//vis2->alpha /= 2;
//vis2->RenderStyle = STYLE_Add;
if (vis->idepth < ds->siz1)
{ // Left is in back, right is in front
vis->sector = ds->curline->backsector;
vis2->sector = ds->curline->frontsector;
}
else
{ // Right is in front, left is in back
vis->sector = ds->curline->frontsector;
vis2->sector = ds->curline->backsector;
}
}
}
}
while (dseg < numdrawsegs && ++spr < numsprites);
// Repeat for any new sprites that were added.
}
while (start = stop, stop != vissprite_p - vissprites);
}
#endif
#ifdef __GNUC__
static void swap(vissprite_t *&a, vissprite_t *&b)
{
vissprite_t *t = a;
a = b;
b = t;
}
#endif
void R_SortVisSprites (bool (*compare)(vissprite_t *, vissprite_t *), size_t first)
{
int i;
vissprite_t **spr;
vsprcount = int(vissprite_p - &vissprites[first]);
if (vsprcount == 0)
return;
if (spritesortersize < MaxVisSprites)
{
if (spritesorter != NULL)
delete[] spritesorter;
spritesorter = new vissprite_t *[MaxVisSprites];
spritesortersize = MaxVisSprites;
}
if (!(i_compatflags & COMPATF_SPRITESORT))
{
for (i = 0, spr = firstvissprite; i < vsprcount; i++, spr++)
{
spritesorter[i] = *spr;
}
}
else
{
// If the compatibility option is on sprites of equal distance need to
// be sorted in inverse order. This is most easily achieved by
// filling the sort array backwards before the sort.
for (i = 0, spr = firstvissprite + vsprcount-1; i < vsprcount; i++, spr--)
{
spritesorter[i] = *spr;
}
}
std::stable_sort(&spritesorter[0], &spritesorter[vsprcount], compare);
}
//
// R_DrawSprite
//
void R_DrawSprite (vissprite_t *spr)
{
static short clipbot[MAXWIDTH];
static short cliptop[MAXWIDTH];
drawseg_t *ds;
int i;
int r1, r2;
short topclip, botclip;
short *clip1, *clip2;
// [RH] Check for particles
if (spr->pic == NULL)
{
R_DrawParticle (spr);
return;
}
// [RH] Quickly reject sprites with bad x ranges.
if (spr->x1 > spr->x2)
return;
// [RH] Sprites split behind a one-sided line can also be discarded.
if (spr->sector == NULL)
return;
// [RH] Initialize the clipping arrays to their largest possible range
// instead of using a special "not clipped" value. This eliminates
// visual anomalies when looking down and should be faster, too.
topclip = 0;
botclip = viewheight;
// killough 3/27/98:
// Clip the sprite against deep water and/or fake ceilings.
// [RH] rewrote this to be based on which part of the sector is really visible
fixed_t scale = MulScale19 (InvZtoScale, spr->idepth);
if (spr->heightsec &&
!(spr->heightsec->MoreFlags & SECF_IGNOREHEIGHTSEC))
{ // only things in specially marked sectors
if (spr->FakeFlatStat != FAKED_AboveCeiling)
{
fixed_t h = spr->heightsec->floorplane.ZatPoint (spr->gx, spr->gy);
//h = (centeryfrac - FixedMul (h-viewz, spr->yscale)) >> FRACBITS;
h = (centeryfrac - FixedMul (h-viewz, scale)) >> FRACBITS;
if (spr->FakeFlatStat == FAKED_BelowFloor)
{ // seen below floor: clip top
if (h > topclip)
{
topclip = MIN<short> (h, viewheight);
}
}
else
{ // seen in the middle: clip bottom
if (h < botclip)
{
botclip = MAX<short> (0, h);
}
}
}
if (spr->FakeFlatStat != FAKED_BelowFloor)
{
fixed_t h = spr->heightsec->ceilingplane.ZatPoint (spr->gx, spr->gy);
h = (centeryfrac - FixedMul (h-viewz, scale)) >> FRACBITS;
if (spr->FakeFlatStat == FAKED_AboveCeiling)
{ // seen above ceiling: clip bottom
if (h < botclip)
{
botclip = MAX<short> (0, h);
}
}
else
{ // seen in the middle: clip top
if (h > topclip)
{
topclip = MIN<short> (h, viewheight);
}
}
}
}
// killough 3/27/98: end special clipping for deep water / fake ceilings
else if (spr->floorclip)
{ // [RH] Move floorclip stuff from R_DrawVisSprite to here
int clip = ((centeryfrac - FixedMul (spr->texturemid -
(spr->pic->GetHeight() << FRACBITS) +
spr->floorclip, spr->yscale)) >> FRACBITS);
if (clip < botclip)
{
botclip = MAX<short> (0, clip);
}
}
#if 0
// [RH] Sprites that were split by a drawseg should also be clipped
// by the sector's floor and ceiling. (Not sure how/if to handle this
// with fake floors, since those already do clipping.)
if (spr->bSplitSprite &&
(spr->heightsec == NULL || (spr->heightsec->MoreFlags & SECF_IGNOREHEIGHTSEC)))
{
fixed_t h = spr->sector->floorplane.ZatPoint (spr->gx, spr->gy);
h = (centeryfrac - FixedMul (h-viewz, scale)) >> FRACBITS;
if (h < botclip)
{
botclip = MAX<short> (0, h);
}
h = spr->sector->ceilingplane.ZatPoint (spr->gx, spr->gy);
h = (centeryfrac - FixedMul (h-viewz, scale)) >> FRACBITS;
if (h > topclip)
{
topclip = MIN<short> (h, viewheight);
}
}
#endif
i = spr->x2 - spr->x1 + 1;
clip1 = clipbot + spr->x1;
clip2 = cliptop + spr->x1;
do
{
*clip1++ = botclip;
*clip2++ = topclip;
} while (--i);
// Scan drawsegs from end to start for obscuring segs.
// The first drawseg that is closer than the sprite is the clip seg.
// Modified by Lee Killough:
// (pointer check was originally nonportable
// and buggy, by going past LEFT end of array):
// for (ds=ds_p-1 ; ds >= drawsegs ; ds--) old buggy code
for (ds = ds_p; ds-- > firstdrawseg; ) // new -- killough
{
// determine if the drawseg obscures the sprite
if (ds->x1 > spr->x2 || ds->x2 < spr->x1 ||
(!(ds->silhouette & SIL_BOTH) && ds->maskedtexturecol == -1 &&
!ds->bFogBoundary) )
{
// does not cover sprite
continue;
}
r1 = MAX<int> (ds->x1, spr->x1);
r2 = MIN<int> (ds->x2, spr->x2);
fixed_t neardepth, fardepth;
if (ds->sz1 < ds->sz2)
{
neardepth = ds->sz1, fardepth = ds->sz2;
}
else
{
neardepth = ds->sz2, fardepth = ds->sz1;
}
if (neardepth > spr->depth || (fardepth > spr->depth &&
// Check if sprite is in front of draw seg:
DMulScale24 (spr->depth - ds->cy, ds->cdx, ds->cdy, ds->cx - spr->cx) < 0))
{
// seg is behind sprite, so draw the mid texture if it has one
if (ds->maskedtexturecol != -1 || ds->bFogBoundary)
R_RenderMaskedSegRange (ds, r1, r2);
continue;
}
// clip this piece of the sprite
// killough 3/27/98: optimized and made much shorter
// [RH] Optimized further (at least for VC++;
// other compilers should be at least as good as before)
if (ds->silhouette & SIL_BOTTOM) //bottom sil
{
clip1 = clipbot + r1;
clip2 = openings + ds->sprbottomclip + r1 - ds->x1;
i = r2 - r1 + 1;
do
{
if (*clip1 > *clip2)
*clip1 = *clip2;
clip1++;
clip2++;
} while (--i);
}
if (ds->silhouette & SIL_TOP) // top sil
{
clip1 = cliptop + r1;
clip2 = openings + ds->sprtopclip + r1 - ds->x1;
i = r2 - r1 + 1;
do
{
if (*clip1 < *clip2)
*clip1 = *clip2;
clip1++;
clip2++;
} while (--i);
}
}
// all clipping has been performed, so draw the sprite
mfloorclip = clipbot;
mceilingclip = cliptop;
R_DrawVisSprite (spr);
}
//
// R_DrawMasked
//
void R_DrawMasked (void)
{
drawseg_t *ds;
int i;
#if 0
R_SplitVisSprites ();
#endif
R_SortVisSprites (sv_compare, firstvissprite - vissprites);
for (i = vsprcount; i > 0; i--)
{
R_DrawSprite (spritesorter[i-1]);
}
// render any remaining masked mid textures
// Modified by Lee Killough:
// (pointer check was originally nonportable
// and buggy, by going past LEFT end of array):
// for (ds=ds_p-1 ; ds >= drawsegs ; ds--) old buggy code
for (ds = ds_p; ds-- > firstdrawseg; ) // new -- killough
{
if (ds->maskedtexturecol != -1 || ds->bFogBoundary)
{
R_RenderMaskedSegRange (ds, ds->x1, ds->x2);
}
}
// draw the psprites on top of everything but does not draw on side views
if (!viewangleoffset)
{
R_DrawPlayerSprites ();
}
}
//
// [RH] Particle functions
//
// [BC] Allow the maximum number of particles to be specified by a cvar (so people
// with lots of nice hardware can have lots of particles!).
CUSTOM_CVAR( Int, r_maxparticles, 4000, CVAR_ARCHIVE )
{
if ( self == 0 )
self = 4000;
else if ( self < 100 )
self = 100;
if ( gamestate != GS_STARTUP )
{
R_DeinitParticles( );
R_InitParticles( );
}
}
void R_InitParticles ()
{
const char *i;
if ((i = Args->CheckValue ("-numparticles")))
NumParticles = atoi (i);
// [BC] Use r_maxparticles now.
else
NumParticles = r_maxparticles;
// This should be good, but eh...
if ( NumParticles < 100 )
NumParticles = 100;
- Fixed: The names in the Depths array in m_options.cpp were never freed. - Fixed: FDoomEdMap needed a destructor. - Fixed: Decal animators were never freed. - Fixed: Colormaps were never freed. - Fixed: Memory allocated in R_InitTranslationTables() was never freed. - Fixed: R_InitParticles() allocated way more memory than it needed to. (And the particle memory was never freed, either.) - Fixed: FMetaTable::FreeMeta() should use delete[] to free string metadata. - Fixed: FConfigFile::ClearCurrentSection() must cast the entry to a char * before deleting it, because that's the way it was allocated. - Fixed definitions of DeadZombieMan and DeadShotgunGuy in doom/deadthings.txt. Skip_super resets the dropitem list, so having it after "DropItem None" is pointless. - Fixed: Decorate DropItem information was never freed. - Fixed: FinishStates() allocated even 0-entry state arrays. - Fixed: Default actor instances were never freed. - Fixed: FRandomSoundList never freed its sound list. - Fixed: Level and cluster strings read from MAPINFO were never freed. - Fixed: Episode names were never freed. - Fixed: InverseColormap and GoldColormap were never freed. Since they're always allocated, they can just be arrays rather than pointers. - Fixed: FFont destructor never freed any of the character data or the font's name. - Fixed: Fonts were not freed at exit. - Fixed: FStringTable::LoadLanguage() did not call SC_Close(). - Fixed: When using the -iwad parameter, IdentifyVersion() did not release the buffer it created to hold the parameter's path. SVN r88 (trunk)
2006-05-09 03:40:15 +00:00
Particles = new particle_t[NumParticles];
R_ClearParticles ();
atterm (R_DeinitParticles);
}
void R_DeinitParticles()
{
if (Particles != NULL)
{
delete[] Particles;
Particles = NULL;
}
}
void R_ClearParticles ()
{
int i;
memset (Particles, 0, NumParticles * sizeof(particle_t));
ActiveParticles = NO_PARTICLE;
InactiveParticles = 0;
for (i = 0; i < NumParticles-1; i++)
Particles[i].tnext = i + 1;
Particles[i].tnext = NO_PARTICLE;
}
// Group particles by subsectors. Because particles are always
// in motion, there is little benefit to caching this information
// from one frame to the next.
void R_FindParticleSubsectors ()
{
if (ParticlesInSubsec.Size() < (size_t)numsubsectors)
{
ParticlesInSubsec.Reserve (numsubsectors - ParticlesInSubsec.Size());
}
clearbufshort (&ParticlesInSubsec[0], numsubsectors, NO_PARTICLE);
if (!r_particles)
{
return;
}
for (WORD i = ActiveParticles; i != NO_PARTICLE; i = Particles[i].tnext)
{
subsector_t *ssec = R_PointInSubsector (Particles[i].x, Particles[i].y);
int ssnum = int(ssec-subsectors);
Particles[i].subsector = ssec;
Particles[i].snext = ParticlesInSubsec[ssnum];
ParticlesInSubsec[ssnum] = i;
}
}
void R_ProjectParticle (particle_t *particle, const sector_t *sector, int shade, int fakeside)
{
fixed_t tr_x;
fixed_t tr_y;
fixed_t tx, ty;
fixed_t tz, tiz;
fixed_t xscale, yscale;
int x1, x2, y1, y2;
vissprite_t* vis;
sector_t* heightsec = NULL;
BYTE* map;
// transform the origin point
tr_x = particle->x - viewx;
tr_y = particle->y - viewy;
tz = DMulScale20 (tr_x, viewtancos, tr_y, viewtansin);
// particle is behind view plane?
if (tz < MINZ)
return;
tx = DMulScale20 (tr_x, viewsin, -tr_y, viewcos);
// Flip for mirrors
if (MirrorFlags & RF_XFLIP)
{
tx = viewwidth - tx - 1;
}
// too far off the side?
if (tz <= abs (tx))
return;
tiz = 268435456 / tz;
xscale = centerx * tiz;
// calculate edges of the shape
int psize = particle->size << (12-3);
x1 = MAX<int> (WindowLeft, (centerxfrac + MulScale12 (tx-psize, xscale)) >> FRACBITS);
x2 = MIN<int> (WindowRight, (centerxfrac + MulScale12 (tx+psize, xscale)) >> FRACBITS);
if (x1 >= x2)
return;
yscale = MulScale16 (yaspectmul, xscale);
ty = particle->z - viewz;
psize <<= 4;
y1 = (centeryfrac - FixedMul (ty+psize, yscale)) >> FRACBITS;
y2 = (centeryfrac - FixedMul (ty-psize, yscale)) >> FRACBITS;
// Clip the particle now. Because it's a point and projected as its subsector is
// entered, we don't need to clip it to drawsegs like a normal sprite.
// Clip particles behind walls.
if (y1 < ceilingclip[x1]) y1 = ceilingclip[x1];
if (y1 < ceilingclip[x2-1]) y1 = ceilingclip[x2-1];
if (y2 >= floorclip[x1]) y2 = floorclip[x1] - 1;
if (y2 >= floorclip[x2-1]) y2 = floorclip[x2-1] - 1;
if (y1 > y2)
return;
// Clip particles above the ceiling or below the floor.
heightsec = sector->GetHeightSec();
const secplane_t *topplane;
const secplane_t *botplane;
FTextureID toppic;
FTextureID botpic;
if (heightsec) // only clip things which are in special sectors
{
if (fakeside == FAKED_AboveCeiling)
{
topplane = &sector->ceilingplane;
botplane = &heightsec->ceilingplane;
toppic = sector->GetTexture(sector_t::ceiling);
botpic = heightsec->GetTexture(sector_t::ceiling);
map = heightsec->ColorMap->Maps;
}
else if (fakeside == FAKED_BelowFloor)
{
topplane = &heightsec->floorplane;
botplane = &sector->floorplane;
toppic = heightsec->GetTexture(sector_t::floor);
botpic = sector->GetTexture(sector_t::floor);
map = heightsec->ColorMap->Maps;
}
else
{
topplane = &heightsec->ceilingplane;
botplane = &heightsec->floorplane;
toppic = heightsec->GetTexture(sector_t::ceiling);
botpic = heightsec->GetTexture(sector_t::floor);
map = sector->ColorMap->Maps;
}
}
else
{
topplane = &sector->ceilingplane;
botplane = &sector->floorplane;
toppic = sector->GetTexture(sector_t::ceiling);
botpic = sector->GetTexture(sector_t::floor);
map = sector->ColorMap->Maps;
}
if (botpic != skyflatnum && particle->z < botplane->ZatPoint (particle->x, particle->y))
return;
if (toppic != skyflatnum && particle->z >= topplane->ZatPoint (particle->x, particle->y))
return;
// store information in a vissprite
vis = R_NewVisSprite ();
vis->heightsec = heightsec;
vis->xscale = xscale;
// vis->yscale = FixedMul (xscale, InvZtoScale);
vis->yscale = xscale;
vis->depth = tz;
vis->idepth = (DWORD)DivScale32 (1, tz) >> 1;
vis->cx = tx;
vis->gx = particle->x;
vis->gy = particle->y;
vis->gz = y1;
vis->gzt = y2;
vis->x1 = x1;
vis->x2 = x2;
vis->Translation = 0;
vis->startfrac = particle->color;
vis->pic = NULL;
vis->renderflags = particle->trans;
vis->FakeFlatStat = fakeside;
vis->floorclip = 0;
vis->heightsec = heightsec;
if (fixedlightlev >= 0)
{
vis->colormap = map + fixedlightlev;
}
else if (fixedcolormap)
{
vis->colormap = fixedcolormap;
}
else
{
// Using MulScale15 instead of 16 makes particles slightly more visible
// than regular sprites.
vis->colormap = map + (GETPALOOKUP (MulScale15 (tiz, r_SpriteVisibility),
shade) << COLORMAPSHIFT);
}
}
static void R_DrawMaskedSegsBehindParticle (const vissprite_t *vis)
{
const int x1 = vis->x1;
const int x2 = vis->x2;
// Draw any masked textures behind this particle so that when the
// particle is drawn, it will be in front of them.
for (unsigned int p = InterestingDrawsegs.Size(); p-- > FirstInterestingDrawseg; )
{
drawseg_t *ds = &drawsegs[InterestingDrawsegs[p]];
if (ds->x1 >= x2 || ds->x2 < x1)
{
continue;
}
if (Scale (ds->siz2 - ds->siz1, (x2 + x1)/2 - ds->sx1, ds->sx2 - ds->sx1) + ds->siz1 < vis->idepth)
{
R_RenderMaskedSegRange (ds, MAX<int> (ds->x1, x1), MIN<int> (ds->x2, x2-1));
}
}
}
void R_DrawParticle (vissprite_t *vis)
{
DWORD *bg2rgb;
int spacing;
BYTE *dest;
DWORD fg;
BYTE color = vis->colormap[vis->startfrac];
int yl = vis->gz;
int ycount = vis->gzt - yl + 1;
int x1 = vis->x1;
int countbase = vis->x2 - x1 + 1;
R_DrawMaskedSegsBehindParticle (vis);
// vis->renderflags holds translucency level (0-255)
{
fixed_t fglevel, bglevel;
DWORD *fg2rgb;
fglevel = ((vis->renderflags + 1) << 8) & ~0x3ff;
bglevel = FRACUNIT-fglevel;
fg2rgb = Col2RGB8[fglevel>>10];
bg2rgb = Col2RGB8[bglevel>>10];
fg = fg2rgb[color];
}
- Ported vlinetallasm4 to AMD64 assembly. Even with the increased number of registers AMD64 provides, this routine still needs to be written as self- modifying code for maximum performance. The additional registers do allow for further optimization over the x86 version by allowing all four pixels to be in flight at the same time. The end result is that AMD64 ASM is about 2.18 times faster than AMD64 C and about 1.06 times faster than x86 ASM. (For further comparison, AMD64 C and x86 C are practically the same for this function.) Should I port any more assembly to AMD64, mvlineasm4 is the most likely candidate, but it's not used enough at this point to bother. Also, this may or may not work with Linux at the moment, since it doesn't have the eh_handler metadata. Win64 is easier, since I just need to structure the function prologue and epilogue properly and use some assembler directives/macros to automatically generate the metadata. And that brings up another point: You need YASM to assemble the AMD64 code, because NASM doesn't support the Win64 metadata directives. - Added an SSE version of DoBlending. This is strictly C intrinsics. VC++ still throws around unneccessary register moves. GCC seems to be pretty close to optimal, requiring only about 2 cycles/color. They're both faster than my hand-written MMX routine, so I don't need to feel bad about not hand-optimizing this for x64 builds. - Removed an extra instruction from DoBlending_MMX, transposed two instructions, and unrolled it once, shaving off about 80 cycles from the time required to blend 256 palette entries. Why? Because I tried writing a C version of the routine using compiler intrinsics and was appalled by all the extra movq's VC++ added to the code. GCC was better, but still generated extra instructions. I only wanted a C version because I can't use inline assembly with VC++'s x64 compiler, and x64 assembly is a bit of a pain. (It's a pain because Linux and Windows have different calling conventions, and you need to maintain extra metadata for functions.) So, the assembly version stays and the C version stays out. - Removed all the pixel doubling r_detail modes, since the one platform they were intended to assist (486) actually sees very little benefit from them. - Rewrote CheckMMX in C and renamed it to CheckCPU. - Fixed: CPUID function 0x80000005 is specified to return detailed L1 cache only for AMD processors, so we must not use it on other architectures, or we end up overwriting the L1 cache line size with 0 or some other number we don't actually understand. SVN r1134 (trunk)
2008-08-09 03:13:43 +00:00
spacing = RenderTarget->GetPitch() - countbase;
dest = ylookup[yl] + x1 + dc_destorg;
do
{
int count = countbase;
do
{
DWORD bg = bg2rgb[*dest];
bg = (fg+bg) | 0x1f07c1f;
*dest++ = RGB32k[0][0][bg & (bg>>15)];
} while (--count);
dest += spacing;
} while (--ycount);
}