gzdoom/src/win32/fb_d3d9.cpp

1840 lines
44 KiB
C++
Raw Normal View History

2006-11-19 02:10:25 +00:00
/*
** fb_d3d9.cpp
** Code to let ZDoom use Direct3D 9 as a simple framebuffer
**
**---------------------------------------------------------------------------
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
** Copyright 1998-2008 Randy Heit
2006-11-19 02:10:25 +00:00
** All rights reserved.
**
** Redistribution and use in source and binary forms, with or without
** modification, are permitted provided that the following conditions
** are met:
**
** 1. Redistributions of source code must retain the above copyright
** notice, this list of conditions and the following disclaimer.
** 2. Redistributions in binary form must reproduce the above copyright
** notice, this list of conditions and the following disclaimer in the
** documentation and/or other materials provided with the distribution.
** 3. The name of the author may not be used to endorse or promote products
** derived from this software without specific prior written permission.
**
** THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
** IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
** OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
** IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
** INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
** NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
** DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
** THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
** THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
**---------------------------------------------------------------------------
**
** This file does _not_ implement hardware-acclerated rendering. It is just
** a means of getting the pixel data to the screen in a more reliable
** method on modern hardware by copying the entire frame to a texture,
** drawing that to the screen, and presenting.
*/
// HEADER FILES ------------------------------------------------------------
#ifdef _DEBUG
#define D3D_DEBUG_INFO
#endif
#define DIRECT3D_VERSION 0x0900
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#include <d3d9.h>
#include <stdio.h>
#define USE_WINDOWS_DWORD
#include "doomtype.h"
#include "c_dispatch.h"
#include "templates.h"
#include "i_system.h"
#include "i_video.h"
Note: I have not tried compiling these recent changes under Linux. I wouldn't be surprised if it doesn't work. - Reorganized the network startup loops so now they are event driven. There is a single function that gets called to drive it, and it uses callbacks to perform the different stages of the synchronization. This lets me have a nice, responsive abort button instead of the previous unannounced hit-escape-to- abort behavior, and I think the rearranged code is slightly easier to understand too. - Increased the number of bytes for version info during D_ArbitrateNetStart(), in preparation for the day when NETGAMEVERSION requires more than one byte. - I noticed an issue with Vista RC1 and the new fatal error setup. Even after releasing a DirectDraw or Direct3D interface, the DWM can still use the last image drawn using them when it composites the window. It doesn't always do it but it does often enough that it is a real problem. At this point, I don't know if it's a problem with the release version of Vista or not. After messing around, I discovered the problem was caused by ~Win32Video() hiding the window and then having it immediately shown soon after. The DWM kept an image of the window to do the transition effect with, and then when it didn't get a chance to do the transition, it didn't properly forget about its saved image and kept plastering it on top of everything else underneath. - Added a network synchronization panel to the window during netgame startup. - Fixed: PClass::CreateDerivedClass() must initialize StateList to NULL. Otherwise, classic DECORATE definitions generate a big, fat crash. - Resurrected the R_Init progress bar, now as a standard Windows control. - Removed the sound failure dialog. The FMOD setup already defaulted to no sound if initialization failed, so this only applies when snd_output is set to "alternate" which now also falls back to no sound. In addition, it wasn't working right, and I didn't feel like fixing it for the probably 0% of users it affected. - Fixed: The edit control used for logging output added text in reverse order on Win9x. - Went back to the roots and made graphics initialization one of the last things to happen during setup. Now the startup text is visible again. More importantly, the main window is no longer created invisible, which seems to cause trouble with it not always appearing in the taskbar. The fatal error dialog is now also embedded in the main window instead of being a separate modal dialog, so you can play with the log window to see any problems that might be reported there. Rather than completely restoring the original startup order, I tried to keep things as close to the way they were with early graphics startup. In particular, V_Init() now creates a dummy screen so that things that need screen dimensions can get them. It gets replaced by the real screen later in I_InitGraphics(). Will need to check this under Linux to make sure it didn't cause any problems there. - Removed the following stubs that just called functions in Video: - I_StartModeIterator() - I_NextMode() - I_DisplayType() I_FullscreenChanged() was also removed, and a new fullscreen parameter was added to IVideo::StartModeIterator(), since that's all it controlled. - Renamed I_InitHardware() back to I_InitGraphics(), since that's all it's initialized post-1.22. SVN r416 (trunk)
2006-12-19 04:09:10 +00:00
#include "i_input.h"
2006-11-19 02:10:25 +00:00
#include "v_video.h"
#include "v_pfx.h"
#include "stats.h"
#include "doomerrors.h"
- 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
#include "r_draw.h"
#include "r_translate.h"
2006-11-19 02:10:25 +00:00
#include "win32iface.h"
#include <mmsystem.h>
// MACROS ------------------------------------------------------------------
// The number of vertices the vertex buffer should hold.
#define NUM_VERTS 28
2006-11-19 02:10:25 +00:00
// TYPES -------------------------------------------------------------------
IMPLEMENT_CLASS(D3DFB)
struct FBVERTEX
{
FLOAT x, y, z, rhw;
FLOAT tu, tv;
};
#define D3DFVF_FBVERTEX (D3DFVF_XYZRHW|D3DFVF_TEX1)
class D3DTex : public FNativeTexture
{
public:
D3DTex(FTexture *tex, D3DFB *fb);
~D3DTex();
FTexture *GameTex;
IDirect3DTexture9 *Tex;
D3DTex **Prev;
D3DTex *Next;
// Texture coordinates to use for the lower-right corner, should this
// texture prove to be larger than the game texture it represents.
FLOAT TX, TY;
bool IsGray;
bool Create(IDirect3DDevice9 *D3DDevice);
bool Update();
D3DFORMAT GetTexFormat();
FTextureFormat ToTexFmt(D3DFORMAT fmt);
};
- 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
class D3DPal : public FNativeTexture
{
public:
D3DPal(FRemapTable *remap, D3DFB *fb);
~D3DPal();
D3DPal **Prev;
D3DPal *Next;
- 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
IDirect3DTexture9 *Tex;
bool Update();
FRemapTable *Remap;
int RoundedPaletteSize;
};
Note: I have not tried compiling these recent changes under Linux. I wouldn't be surprised if it doesn't work. - Reorganized the network startup loops so now they are event driven. There is a single function that gets called to drive it, and it uses callbacks to perform the different stages of the synchronization. This lets me have a nice, responsive abort button instead of the previous unannounced hit-escape-to- abort behavior, and I think the rearranged code is slightly easier to understand too. - Increased the number of bytes for version info during D_ArbitrateNetStart(), in preparation for the day when NETGAMEVERSION requires more than one byte. - I noticed an issue with Vista RC1 and the new fatal error setup. Even after releasing a DirectDraw or Direct3D interface, the DWM can still use the last image drawn using them when it composites the window. It doesn't always do it but it does often enough that it is a real problem. At this point, I don't know if it's a problem with the release version of Vista or not. After messing around, I discovered the problem was caused by ~Win32Video() hiding the window and then having it immediately shown soon after. The DWM kept an image of the window to do the transition effect with, and then when it didn't get a chance to do the transition, it didn't properly forget about its saved image and kept plastering it on top of everything else underneath. - Added a network synchronization panel to the window during netgame startup. - Fixed: PClass::CreateDerivedClass() must initialize StateList to NULL. Otherwise, classic DECORATE definitions generate a big, fat crash. - Resurrected the R_Init progress bar, now as a standard Windows control. - Removed the sound failure dialog. The FMOD setup already defaulted to no sound if initialization failed, so this only applies when snd_output is set to "alternate" which now also falls back to no sound. In addition, it wasn't working right, and I didn't feel like fixing it for the probably 0% of users it affected. - Fixed: The edit control used for logging output added text in reverse order on Win9x. - Went back to the roots and made graphics initialization one of the last things to happen during setup. Now the startup text is visible again. More importantly, the main window is no longer created invisible, which seems to cause trouble with it not always appearing in the taskbar. The fatal error dialog is now also embedded in the main window instead of being a separate modal dialog, so you can play with the log window to see any problems that might be reported there. Rather than completely restoring the original startup order, I tried to keep things as close to the way they were with early graphics startup. In particular, V_Init() now creates a dummy screen so that things that need screen dimensions can get them. It gets replaced by the real screen later in I_InitGraphics(). Will need to check this under Linux to make sure it didn't cause any problems there. - Removed the following stubs that just called functions in Video: - I_StartModeIterator() - I_NextMode() - I_DisplayType() I_FullscreenChanged() was also removed, and a new fullscreen parameter was added to IVideo::StartModeIterator(), since that's all it controlled. - Renamed I_InitHardware() back to I_InitGraphics(), since that's all it's initialized post-1.22. SVN r416 (trunk)
2006-12-19 04:09:10 +00:00
// EXTERNAL FUNCTION PROTOTYPES --------------------------------------------
2006-11-19 02:10:25 +00:00
// PUBLIC FUNCTION PROTOTYPES ----------------------------------------------
void DoBlending (const PalEntry *from, PalEntry *to, int count, int r, int g, int b, int a);
// PRIVATE FUNCTION PROTOTYPES ---------------------------------------------
// EXTERNAL DATA DECLARATIONS ----------------------------------------------
extern HWND Window;
extern IVideo *Video;
extern BOOL AppActive;
extern int SessionState;
extern bool VidResizing;
EXTERN_CVAR (Bool, fullscreen)
EXTERN_CVAR (Float, Gamma)
EXTERN_CVAR (Int, vid_displaybits)
EXTERN_CVAR (Bool, vid_vsync)
EXTERN_CVAR (Float, transsouls)
2006-11-19 02:10:25 +00:00
extern IDirect3D9 *D3D;
extern cycle_t BlitCycles;
// PRIVATE DATA DEFINITIONS ------------------------------------------------
#include "fb_d3d9_shaders.h"
2006-11-19 02:10:25 +00:00
// PUBLIC DATA DEFINITIONS -------------------------------------------------
// CODE --------------------------------------------------------------------
D3DFB::D3DFB (int width, int height, bool fullscreen)
: BaseWinFB (width, height)
{
D3DPRESENT_PARAMETERS d3dpp;
D3DDevice = NULL;
VertexBuffer = NULL;
FBTexture = NULL;
WindowedRenderTexture = NULL;
2006-11-19 02:10:25 +00:00
PaletteTexture = NULL;
StencilPaletteTexture = NULL;
ShadedPaletteTexture = NULL;
2006-11-19 02:10:25 +00:00
PalTexShader = NULL;
PlainShader = NULL;
PlainStencilShader = NULL;
DimShader = NULL;
GammaFixerShader = NULL;
2006-11-19 02:10:25 +00:00
FBFormat = D3DFMT_UNKNOWN;
PalFormat = D3DFMT_UNKNOWN;
VSync = vid_vsync;
BlendingRect.left = 0;
BlendingRect.top = 0;
BlendingRect.right = FBWidth;
BlendingRect.bottom = FBHeight;
UseBlendingRect = false;
In2D = 0;
Palettes = NULL;
Textures = NULL;
Accel2D = true;
2006-11-19 02:10:25 +00:00
Gamma = 1.0;
FlashConstants[0][3] = FlashConstants[0][2] = FlashConstants[0][1] = FlashConstants[0][0] = 0;
FlashConstants[1][3] = FlashConstants[1][2] = FlashConstants[1][1] = FlashConstants[1][0] = 1;
2006-11-19 02:10:25 +00:00
FlashColor = 0;
FlashAmount = 0;
NeedGammaUpdate = false;
NeedPalUpdate = false;
if (MemBuffer == NULL)
{
return;
}
memcpy (SourcePalette, GPalette.BaseColors, sizeof(PalEntry)*256);
Windowed = !(static_cast<Win32Video *>(Video)->GoFullscreen (fullscreen));
TrueHeight = height;
if (fullscreen)
{
for (Win32Video::ModeInfo *mode = static_cast<Win32Video *>(Video)->m_Modes; mode != NULL; mode = mode->next)
{
if (mode->width == Width && mode->height == Height)
{
TrueHeight = mode->realheight;
break;
}
}
}
FillPresentParameters (&d3dpp, fullscreen, VSync);
HRESULT hr;
if (FAILED(hr = D3D->CreateDevice (D3DADAPTER_DEFAULT, D3DDEVTYPE_HAL, Window,
D3DCREATE_SOFTWARE_VERTEXPROCESSING | D3DCREATE_FPU_PRESERVE, &d3dpp, &D3DDevice)))
2006-11-19 02:10:25 +00:00
{
D3DDevice = NULL;
if (fullscreen)
{
d3dpp.BackBufferFormat = D3DFMT_R5G6B5;
if (FAILED(D3D->CreateDevice (D3DADAPTER_DEFAULT, D3DDEVTYPE_HAL, Window,
D3DCREATE_SOFTWARE_VERTEXPROCESSING | D3DCREATE_FPU_PRESERVE, &d3dpp, &D3DDevice)))
2006-11-19 02:10:25 +00:00
{
D3DDevice = NULL;
}
}
}
if (D3DDevice != NULL)
{
CreateResources ();
- 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
// Be sure we know what the alpha blend is
D3DDevice->SetRenderState(D3DRS_ALPHABLENDENABLE, FALSE);
AlphaBlendEnabled = FALSE;
AlphaSrcBlend = D3DBLEND(0);
AlphaDestBlend = D3DBLEND(0);
2006-11-19 02:10:25 +00:00
}
}
D3DFB::~D3DFB ()
{
KillNativeTexs();
KillNativePals();
2006-11-19 02:10:25 +00:00
ReleaseResources ();
if (D3DDevice != NULL)
{
D3DDevice->Release();
}
}
void D3DFB::FillPresentParameters (D3DPRESENT_PARAMETERS *pp, bool fullscreen, bool vsync)
{
memset (pp, 0, sizeof(*pp));
pp->Windowed = !fullscreen;
pp->SwapEffect = D3DSWAPEFFECT_DISCARD;
pp->BackBufferWidth = Width;
pp->BackBufferHeight = TrueHeight;
pp->BackBufferFormat = fullscreen ? D3DFMT_X8R8G8B8 : D3DFMT_UNKNOWN;
pp->hDeviceWindow = Window;
pp->PresentationInterval = vsync ? D3DPRESENT_INTERVAL_ONE : D3DPRESENT_INTERVAL_IMMEDIATE;
}
bool D3DFB::CreateResources ()
{
if (!Windowed)
{
// Remove the window border in fullscreen mode
SetWindowLong (Window, GWL_STYLE, WS_POPUP|WS_VISIBLE|WS_SYSMENU);
2006-11-19 02:10:25 +00:00
}
else
{
// Resize the window to match desired dimensions
int sizew = Width + GetSystemMetrics (SM_CXSIZEFRAME)*2;
int sizeh = Height + GetSystemMetrics (SM_CYSIZEFRAME) * 2 +
GetSystemMetrics (SM_CYCAPTION);
LOG2 ("Resize window to %dx%d\n", sizew, sizeh);
VidResizing = true;
// Make sure the window has a border in windowed mode
SetWindowLong (Window, GWL_STYLE, WS_VISIBLE|WS_OVERLAPPEDWINDOW);
if (GetWindowLong (Window, GWL_EXSTYLE) & WS_EX_TOPMOST)
{
// Direct3D 9 will apparently add WS_EX_TOPMOST to fullscreen windows,
// and removing it is a little tricky. Using SetWindowLongPtr to clear it
// will not do the trick, but sending the window behind everything will.
SetWindowPos (Window, HWND_BOTTOM, 0, 0, sizew, sizeh,
SWP_DRAWFRAME | SWP_NOCOPYBITS | SWP_NOMOVE);
SetWindowPos (Window, HWND_TOP, 0, 0, 0, 0, SWP_NOCOPYBITS | SWP_NOMOVE | SWP_NOSIZE);
}
else
2006-11-19 02:10:25 +00:00
{
SetWindowPos (Window, NULL, 0, 0, sizew, sizeh,
SWP_DRAWFRAME | SWP_NOCOPYBITS | SWP_NOMOVE | SWP_NOZORDER);
2006-11-19 02:10:25 +00:00
}
I_RestoreWindowedPos ();
2006-11-19 02:10:25 +00:00
VidResizing = false;
}
- 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
SM14 = false;
if (FAILED(D3DDevice->CreatePixelShader (PalTexShader20Def, &PalTexShader)) &&
(SM14 = true, FAILED(D3DDevice->CreatePixelShader (PalTexShader14Def, &PalTexShader))))
{
return false;
}
if (FAILED(D3DDevice->CreatePixelShader (PlainShaderDef, &PlainShader)) ||
FAILED(D3DDevice->CreatePixelShader (PlainStencilDef, &PlainStencilShader)) ||
FAILED(D3DDevice->CreatePixelShader (DimShaderDef, &DimShader)))
{
return false;
}
if (FAILED(D3DDevice->CreatePixelShader (GammaFixerDef, &GammaFixerShader)))
{
Printf ("Windowed mode gamma will not work.\n");
GammaFixerShader = NULL;
}
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
CurPixelShader = NULL;
memset(Constant, 0, sizeof(Constant));
if (!CreateFBTexture() ||
!CreatePaletteTexture() ||
!CreateStencilPaletteTexture() ||
!CreateShadedPaletteTexture())
2006-11-19 02:10:25 +00:00
{
return false;
}
if (!CreateVertexes())
{
return false;
}
SetGamma (Gamma);
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
2006-11-19 02:10:25 +00:00
return true;
}
void D3DFB::ReleaseResources ()
{
I_SaveWindowedPos ();
2006-11-19 02:10:25 +00:00
if (FBTexture != NULL)
{
FBTexture->Release();
FBTexture = NULL;
}
if (WindowedRenderTexture != NULL)
{
WindowedRenderTexture->Release();
WindowedRenderTexture = NULL;
}
2006-11-19 02:10:25 +00:00
if (VertexBuffer != NULL)
{
VertexBuffer->Release();
VertexBuffer = NULL;
}
if (PaletteTexture != NULL)
{
PaletteTexture->Release();
PaletteTexture = NULL;
}
if (StencilPaletteTexture != NULL)
{
StencilPaletteTexture->Release();
StencilPaletteTexture = NULL;
}
if (ShadedPaletteTexture != NULL)
{
ShadedPaletteTexture->Release();
ShadedPaletteTexture = NULL;
}
2006-11-19 02:10:25 +00:00
if (PalTexShader != NULL)
{
PalTexShader->Release();
PalTexShader = NULL;
}
if (PlainShader != NULL)
{
PlainShader->Release();
PlainShader = NULL;
}
if (PlainStencilShader != NULL)
{
PlainStencilShader->Release();
PlainStencilShader = NULL;
}
if (DimShader != NULL)
{
DimShader->Release();
DimShader = NULL;
}
if (GammaFixerShader != NULL)
{
GammaFixerShader->Release();
GammaFixerShader = NULL;
}
2006-11-19 02:10:25 +00:00
}
bool D3DFB::Reset ()
{
D3DPRESENT_PARAMETERS d3dpp;
// Free resources created with D3DPOOL_DEFAULT.
if (FBTexture != NULL)
{
FBTexture->Release();
FBTexture = NULL;
}
if (WindowedRenderTexture != NULL)
{
WindowedRenderTexture->Release();
WindowedRenderTexture = NULL;
}
2006-11-19 02:10:25 +00:00
if (VertexBuffer != NULL)
{
VertexBuffer->Release();
VertexBuffer = NULL;
}
FillPresentParameters (&d3dpp, !Windowed, VSync);
if (!SUCCEEDED(D3DDevice->Reset (&d3dpp)))
{
return false;
}
if (!CreateFBTexture() || !CreateVertexes())
{
return false;
}
return true;
}
//==========================================================================
//
// D3DFB :: KillNativePals
//
// Frees all native palettes.
//
//==========================================================================
void D3DFB::KillNativePals()
{
while (Palettes != NULL)
{
delete Palettes;
}
}
//==========================================================================
//
// D3DFB :: KillNativeTexs
//
// Frees all native textures.
//
//==========================================================================
void D3DFB::KillNativeTexs()
{
while (Textures != NULL)
{
delete Textures;
}
}
//==========================================================================
//
// D3DFB :: KillNativeNonPalettedTexs
//
// Frees all native textures that aren't paletted.
//
//==========================================================================
void D3DFB::KillNativeNonPalettedTexs()
{
D3DTex *tex;
D3DTex *next;
for (tex = Textures; tex != NULL; tex = next)
{
next = tex->Next;
if (tex->GetTexFormat() != D3DFMT_L8)
{
delete tex;
}
}
}
2006-11-19 02:10:25 +00:00
bool D3DFB::CreateFBTexture ()
{
if (FAILED(D3DDevice->CreateTexture (Width, Height, 1, D3DUSAGE_DYNAMIC, D3DFMT_L8, D3DPOOL_DEFAULT, &FBTexture, NULL)))
{
int pow2width, pow2height, i;
for (i = 1; i < Width; i <<= 1) {} pow2width = i;
for (i = 1; i < Height; i <<= 1) {} pow2height = i;
if (FAILED(D3DDevice->CreateTexture (pow2width, pow2height, 1, D3DUSAGE_DYNAMIC, D3DFMT_L8, D3DPOOL_DEFAULT, &FBTexture, NULL)))
{
return false;
}
else
{
FBWidth = pow2width;
FBHeight = pow2height;
}
}
else
{
FBWidth = Width;
FBHeight = Height;
}
if (Windowed && GammaFixerShader)
{
if (FAILED(D3DDevice->CreateTexture (FBWidth, FBHeight, 1, D3DUSAGE_RENDERTARGET, D3DFMT_X8R8G8B8, D3DPOOL_DEFAULT, &WindowedRenderTexture, NULL)))
{
WindowedRenderTexture = false;
}
}
2006-11-19 02:10:25 +00:00
return true;
}
bool D3DFB::CreatePaletteTexture ()
{
if (FAILED(D3DDevice->CreateTexture (256, 1, 1, 0, D3DFMT_A8R8G8B8, D3DPOOL_MANAGED, &PaletteTexture, NULL)))
{
return false;
2006-11-19 02:10:25 +00:00
}
PalFormat = D3DFMT_A8R8G8B8;
return true;
}
bool D3DFB::CreateStencilPaletteTexture()
{
// The stencil palette is a special palette where the first entry is zero alpha,
// and everything else is white with full alpha.
if (FAILED(D3DDevice->CreateTexture(256, 1, 1, 0, D3DFMT_A8R8G8B8, D3DPOOL_MANAGED, &StencilPaletteTexture, NULL)))
{
return false;
}
D3DLOCKED_RECT lockrect;
if (SUCCEEDED(StencilPaletteTexture->LockRect(0, &lockrect, NULL, 0)))
2006-11-19 02:10:25 +00:00
{
DWORD *pix = (DWORD *)lockrect.pBits;
*pix = 0;
memset(pix + 1, 0xFF, 255*4);
StencilPaletteTexture->UnlockRect(0);
}
return true;
}
bool D3DFB::CreateShadedPaletteTexture()
{
// The shaded palette is similar to the stencil palette, except each entry's
// alpha is the same as its index.
if (FAILED(D3DDevice->CreateTexture(256, 1, 1, 0, D3DFMT_A8R8G8B8, D3DPOOL_MANAGED, &ShadedPaletteTexture, NULL)))
{
return false;
}
D3DLOCKED_RECT lockrect;
if (SUCCEEDED(ShadedPaletteTexture->LockRect(0, &lockrect, NULL, 0)))
{
BYTE *pix = (BYTE *)lockrect.pBits;
for (int i = 0; i < 256; ++i)
{
pix[3] = i;
pix[2] = pix[1] = pix[0] = 255;
pix += 4;
}
ShadedPaletteTexture->UnlockRect(0);
2006-11-19 02:10:25 +00:00
}
return true;
}
bool D3DFB::CreateVertexes ()
{
if (FAILED(D3DDevice->CreateVertexBuffer (sizeof(FBVERTEX)*NUM_VERTS, D3DUSAGE_WRITEONLY, D3DFVF_FBVERTEX, D3DPOOL_DEFAULT, &VertexBuffer, NULL)) ||
!UploadVertices())
{
return false;
}
return true;
}
bool D3DFB::UploadVertices()
2006-11-19 02:10:25 +00:00
{
float top = (TrueHeight - Height) * 0.5f - 0.5f;
float right = float(Width) - 0.5f;
2007-01-09 04:40:58 +00:00
float bot = float(Height) + top;
2006-11-19 02:10:25 +00:00
float texright = float(Width) / float(FBWidth);
float texbot = float(Height) / float(FBHeight);
void *pverts;
if ((BlendingRect.left <= 0 && BlendingRect.right >= FBWidth &&
BlendingRect.top <= 0 && BlendingRect.bottom >= FBHeight) ||
BlendingRect.left >= BlendingRect.right ||
BlendingRect.top >= BlendingRect.bottom)
2006-11-19 02:10:25 +00:00
{
// Blending rect covers the whole screen, so only need 4 verts.
FBVERTEX verts[4] =
{
{ -0.5f, top, 0.5f, 1.f, 0.f, 0.f },
{ right, top, 0.5f, 1.f, texright, 0.f },
{ right, bot, 0.5f, 1.f, texright, texbot },
{ -0.5f, bot, 0.5f, 1.f, 0.f, texbot }
};
if (SUCCEEDED(VertexBuffer->Lock(0, sizeof(verts), &pverts, 0)))
{
memcpy (pverts, verts, sizeof(verts));
VertexBuffer->Unlock();
return true;
}
2006-11-19 02:10:25 +00:00
return false;
}
// Only the 3D area of the screen is effected by palette flashes.
// So we create some boxes around it that can be drawn without the
// flash. These include the corners of the view area so I can be
// sure the texture interpolation is consistant. (Well, actually,
// since it's a 1-to-1 pixel mapping, it shouldn't matter.)
float mxl = float(BlendingRect.left) - 0.5f;
float mxr = float(BlendingRect.right) - 0.5f;
float myt = float(BlendingRect.top) + top;
float myb = float(BlendingRect.bottom) + top;
float tmxl = float(BlendingRect.left) / float(Width) * texright;
float tmxr = float(BlendingRect.right) / float(Width) * texright;
float tmyt = float(BlendingRect.top) / float(Height) * texbot;
float tmyb = float(BlendingRect.bottom) / float(Height) * texbot;
/* +-------------------+
| |
+-----+-------+-----+
| | | |
| | | |
| | | |
+-----+-------+-----+
| |
+-------------------+ */
FBVERTEX verts[28] =
{
// The whole screen, for when no blending is happening
{ -0.5f, top, 0.5f, 1.f, 0.f, 0.f }, // 0
{ right, top, 0.5f, 1.f, texright, 0.f },
{ right, bot, 0.5f, 1.f, texright, texbot },
{ -0.5f, bot, 0.5f, 1.f, 0.f, texbot },
// Left area
{ -0.5f, myt, 0.5f, 1.f, 0.f, tmyt }, // 4
{ mxl, myt, 0.5f, 1.f, tmxl, tmyt },
{ mxl, myb, 0.5f, 1.f, tmxl, tmyb },
{ -0.5f, myb, 0.5f, 1.f, 0.f, tmyb },
// Right area
{ mxr, myt, 0.5f, 1.f, tmxr, tmyt }, // 8
{ right, myt, 0.5f, 1.f, texright, tmyt },
{ right, myb, 0.5f, 1.f, texright, tmyb },
{ mxr, myb, 0.5f, 1.f, tmxr, tmyb },
// Bottom area
{ -0.5f, bot, 0.5f, 1.f, 0.f, texbot }, // 12
{ -0.5f, myb, 0.5f, 1.f, 0.f, tmyb },
{ mxl, myb, 0.5f, 1.f, tmxl, tmyb },
{ mxr, myb, 0.5f, 1.f, tmxr, tmyb },
{ right, myb, 0.5f, 1.f, texright, tmyb },
{ right, bot, 0.5f, 1.f, texright, texbot },
// Top area
{ right, top, 0.5f, 1.f, texright, 0.f }, // 18
{ right, myt, 0.5f, 1.f, texright, tmyt },
{ mxr, myt, 0.5f, 1.f, tmxr, tmyt },
{ mxl, myt, 0.5f, 1.f, tmxl, tmyt },
{ -0.5f, myt, 0.5f, 1.f, 0.f, tmyt },
{ -0.5f, top, 0.5f, 1.f, 0.f, 0.f },
// Middle (blended) area
{ mxl, myt, 0.5f, 1.f, tmxl, tmyt }, // 24
{ mxr, myt, 0.5f, 1.f, tmxr, tmyt },
{ mxr, myb, 0.5f, 1.f, tmxr, tmyb },
{ mxl, myb, 0.5f, 1.f, tmxl, tmyb }
};
if (SUCCEEDED(VertexBuffer->Lock(0, sizeof(verts), &pverts, 0)))
2006-11-19 02:10:25 +00:00
{
memcpy (pverts, verts, sizeof(verts));
VertexBuffer->Unlock();
return true;
2006-11-19 02:10:25 +00:00
}
return false;
2006-11-19 02:10:25 +00:00
}
int D3DFB::GetPageCount ()
{
return 1;
}
void D3DFB::PaletteChanged ()
{
}
int D3DFB::QueryNewPalette ()
{
return 0;
}
bool D3DFB::IsValid ()
{
return D3DDevice != NULL;
}
HRESULT D3DFB::GetHR ()
{
return 0;
}
bool D3DFB::IsFullscreen ()
{
return !Windowed;
}
bool D3DFB::Lock ()
{
return Lock(true);
}
bool D3DFB::Lock (bool buffered)
{
if (LockCount++ > 0)
{
return false;
}
Buffer = MemBuffer;
return false;
}
void D3DFB::Unlock ()
{
LOG1 ("Unlock <%d>\n", LockCount);
if (LockCount == 0)
{
return;
}
if (UpdatePending && LockCount == 1)
{
Update ();
}
else if (--LockCount == 0)
{
Buffer = NULL;
}
}
// When In2D == 0: Copy buffer to screen and present
// When In2D == 1: Copy buffer to screen but do not present
// When In2D == 2: Present and set In2D to 0
2006-11-19 02:10:25 +00:00
void D3DFB::Update ()
{
if (In2D == 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
DrawRateStuff();
DoWindowedGamma();
D3DDevice->EndScene();
D3DDevice->Present(NULL, NULL, NULL, NULL);
In2D = 0;
return;
}
2006-11-19 02:10:25 +00:00
if (LockCount != 1)
{
- 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
I_FatalError ("Framebuffer must have exactly 1 lock to be updated");
2006-11-19 02:10:25 +00:00
if (LockCount > 0)
{
UpdatePending = true;
--LockCount;
}
return;
}
- 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
if (In2D == 0)
{
DrawRateStuff();
}
2006-11-19 02:10:25 +00:00
if (NeedGammaUpdate)
{
float psgamma[4];
float igamma;
2006-11-19 02:10:25 +00:00
NeedGammaUpdate = false;
igamma = 1 / Gamma;
if (!Windowed)
{
D3DGAMMARAMP ramp;
for (int i = 0; i < 256; ++i)
{
ramp.blue[i] = ramp.green[i] = ramp.red[i] = WORD(65535.f * powf(i / 255.f, igamma));
}
D3DDevice->SetGammaRamp(0, D3DSGR_CALIBRATE, &ramp);
}
psgamma[2] = psgamma[1] = psgamma[0] = igamma;
psgamma[3] = 1;
D3DDevice->SetPixelShaderConstantF(4, psgamma, 1);
2006-11-19 02:10:25 +00:00
NeedPalUpdate = true;
}
if (NeedPalUpdate)
{
UploadPalette();
}
BlitCycles = 0;
clock (BlitCycles);
LockCount = 0;
PaintToWindow ();
if (In2D == 0)
{
DoWindowedGamma();
D3DDevice->EndScene();
D3DDevice->Present(NULL, NULL, NULL, NULL);
}
2006-11-19 02:10:25 +00:00
unclock (BlitCycles);
LOG1 ("cycles = %d\n", BlitCycles);
Buffer = NULL;
UpdatePending = false;
}
bool D3DFB::PaintToWindow ()
{
HRESULT hr;
if (LockCount != 0)
{
return false;
}
hr = D3DDevice->TestCooperativeLevel();
if (FAILED(hr))
{
if (hr != D3DERR_DEVICENOTRESET || !Reset())
{
Sleep (1);
return false;
}
}
Draw3DPart();
return true;
}
void D3DFB::Draw3DPart()
{
RECT texrect = { 0, 0, Width, Height };
D3DLOCKED_RECT lockrect;
if ((FBWidth == Width && FBHeight == Height &&
SUCCEEDED(FBTexture->LockRect (0, &lockrect, NULL, D3DLOCK_DISCARD))) ||
SUCCEEDED(FBTexture->LockRect (0, &lockrect, &texrect, 0)))
2006-11-19 02:10:25 +00:00
{
if (lockrect.Pitch == Pitch)
{
memcpy (lockrect.pBits, MemBuffer, Width * Height);
}
else
{
BYTE *dest = (BYTE *)lockrect.pBits;
BYTE *src = MemBuffer;
for (int y = 0; y < Height; y++)
{
memcpy (dest, src, Width);
dest += lockrect.Pitch;
src += Pitch;
}
}
FBTexture->UnlockRect (0);
}
if (TrueHeight != Height)
{
// Letterbox! Draw black top and bottom borders.
int topborder = (TrueHeight - Height) / 2;
D3DRECT rects[2] = { { 0, 0, Width, topborder }, { 0, Height + topborder, Width, TrueHeight } };
D3DDevice->Clear (2, rects, D3DCLEAR_TARGET, D3DCOLOR_XRGB(0,0,0), 1.f, 0);
}
D3DDevice->BeginScene();
OldRenderTarget = NULL;
if (WindowedRenderTexture != NULL)
{
IDirect3DSurface9 *targetsurf;
if (FAILED(WindowedRenderTexture->GetSurfaceLevel(0, &targetsurf)) ||
FAILED(D3DDevice->GetRenderTarget(0, &OldRenderTarget)) ||
FAILED(D3DDevice->SetRenderTarget(0, targetsurf)))
{
// Setting the render target failed.
OldRenderTarget = NULL;
}
}
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
SetTexture (0, FBTexture);
SetPaletteTexture(PaletteTexture, 256);
2006-11-19 02:10:25 +00:00
D3DDevice->SetStreamSource (0, VertexBuffer, 0, sizeof(FBVERTEX));
D3DDevice->SetFVF (D3DFVF_FBVERTEX);
D3DDevice->SetPixelShaderConstantF (0, FlashConstants[0], 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
memcpy(Constant, FlashConstants, sizeof(FlashConstants));
SetAlphaBlend(FALSE);
if (!UseBlendingRect || FlashConstants[1][0] == 1)
{ // The whole screen as a single quad.
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 0, 2);
}
else
{ // The screen split up so that only the 3D view is blended.
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 24, 2); // middle
if (!In2D || 1)
{
// The rest is drawn unblended, so reset the shader constant.
static const float FlashZero[2][4] = { { 0, 0, 0, 0 }, { 1, 1, 1, 1 } };
D3DDevice->SetPixelShaderConstantF (0, FlashZero[0], 2);
memcpy(Constant, FlashZero, sizeof(FlashZero));
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 4, 2); // left
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 8, 2); // right
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 12, 4); // bottom
D3DDevice->DrawPrimitive (D3DPT_TRIANGLEFAN, 18, 4); // top
}
}
}
//==========================================================================
//
// D3DFB :: DoWindowedGamma
//
// Draws the render target texture to the real back buffer using a gamma-
// correcting pixel shader.
//
//==========================================================================
void D3DFB::DoWindowedGamma()
{
if (OldRenderTarget != NULL)
{
D3DDevice->SetRenderTarget(0, OldRenderTarget);
D3DDevice->SetStreamSource(0, VertexBuffer, 0, sizeof(FBVERTEX));
D3DDevice->SetFVF(D3DFVF_FBVERTEX);
SetTexture(0, WindowedRenderTexture);
SetPixelShader(GammaFixerShader);
SetAlphaBlend(FALSE);
D3DDevice->DrawPrimitive(D3DPT_TRIANGLEFAN, 0, 2);
OldRenderTarget = NULL;
}
2006-11-19 02:10:25 +00:00
}
void D3DFB::UploadPalette ()
{
D3DLOCKED_RECT lockrect;
int i;
2006-11-19 02:10:25 +00:00
if (SUCCEEDED(PaletteTexture->LockRect (0, &lockrect, NULL, 0)))
{
BYTE *pix = (BYTE *)lockrect.pBits;
- 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
for (i = 0; i < 256; ++i, pix += 4)
2006-11-19 02:10:25 +00:00
{
pix[0] = SourcePalette[i].b;
pix[1] = SourcePalette[i].g;
pix[2] = SourcePalette[i].r;
pix[3] = (i == 0 ? 0 : 255);
// To let masked textures work, the first palette entry's alpha is 0.
2006-11-19 02:10:25 +00:00
}
PaletteTexture->UnlockRect (0);
}
}
PalEntry *D3DFB::GetPalette ()
{
return SourcePalette;
}
void D3DFB::UpdatePalette ()
{
NeedPalUpdate = true;
}
bool D3DFB::SetGamma (float gamma)
{
LOG1 ("SetGamma %g\n", gamma);
Gamma = gamma;
NeedGammaUpdate = true;
return true;
}
bool D3DFB::SetFlash (PalEntry rgb, int amount)
{
FlashColor = rgb;
FlashAmount = amount;
// Fill in the constants for the pixel shader to do linear interpolation between the palette and the flash:
float r = rgb.r / 255.f, g = rgb.g / 255.f, b = rgb.b / 255.f, a = amount / 256.f;
FlashConstants[0][0] = r * a;
FlashConstants[0][1] = g * a;
FlashConstants[0][2] = b * a;
a = 1 - a;
FlashConstants[1][0] = a;
FlashConstants[1][1] = a;
FlashConstants[1][2] = a;
return true;
}
void D3DFB::GetFlash (PalEntry &rgb, int &amount)
{
rgb = FlashColor;
amount = FlashAmount;
}
void D3DFB::GetFlashedPalette (PalEntry pal[256])
{
memcpy (pal, SourcePalette, 256*sizeof(PalEntry));
if (FlashAmount)
{
DoBlending (pal, pal, 256, FlashColor.r, FlashColor.g, FlashColor.b, FlashAmount);
}
}
void D3DFB::SetVSync (bool vsync)
{
if (VSync != vsync)
{
VSync = vsync;
Reset();
}
}
void D3DFB::Blank ()
{
// Only used by movie player, which isn't working with D3D9 yet.
}
void D3DFB::SetBlendingRect(int x1, int y1, int x2, int y2)
{
if (BlendingRect.left != x1 ||
BlendingRect.top != y1 ||
BlendingRect.right != x2 ||
BlendingRect.bottom != y2)
{
BlendingRect.left = x1;
BlendingRect.top = y1;
BlendingRect.right = x2;
BlendingRect.bottom = y2;
if (UploadVertices())
{
UseBlendingRect = ((x1 > 0 || x2 < FBWidth || y1 > 0 || y2 < FBHeight)
&& BlendingRect.left < BlendingRect.right
&& BlendingRect.top < BlendingRect.bottom);
}
}
}
/**************************************************************************/
/* 2D Stuff */
/**************************************************************************/
//==========================================================================
//
// D3DTex Constructor
//
//==========================================================================
D3DTex::D3DTex(FTexture *tex, D3DFB *fb)
{
// Attach to the texture list for the D3DFB
Next = fb->Textures;
if (Next != NULL)
{
Next->Prev = &Next;
}
Prev = &fb->Textures;
fb->Textures = this;
GameTex = tex;
Tex = NULL;
IsGray = false;
Create(fb->D3DDevice);
}
//==========================================================================
//
// D3DTex Destructor
//
//==========================================================================
D3DTex::~D3DTex()
{
if (Tex != NULL)
{
Tex->Release();
Tex = NULL;
}
// Detach from the texture list
*Prev = Next;
if (Next != NULL)
{
Next->Prev = Prev;
}
// Remove link from the game texture
if (GameTex != NULL)
{
GameTex->Native = NULL;
}
}
//==========================================================================
//
// D3DTex :: Create
//
// Creates an IDirect3DTexture9 for the texture and copies the image data
// to it. Note that unlike FTexture, this image is row-major.
//
//==========================================================================
bool D3DTex::Create(IDirect3DDevice9 *D3DDevice)
{
HRESULT hr;
int w, h;
if (Tex != NULL)
{
Tex->Release();
Tex = NULL;
}
w = GameTex->GetWidth();
h = GameTex->GetHeight();
hr = D3DDevice->CreateTexture(w, h, 1, 0,
GetTexFormat(), D3DPOOL_MANAGED, &Tex, NULL);
if (FAILED(hr))
{ // Try again, using power-of-2 sizes
int i;
for (i = 1; i < w; i <<= 1) {} w = i;
for (i = 1; i < h; i <<= 1) {} h = i;
hr = D3DDevice->CreateTexture(w, h, 1, 0,
GetTexFormat(), D3DPOOL_MANAGED, &Tex, NULL);
if (FAILED(hr))
{
return false;
}
}
if (!Update())
{
Tex->Release();
Tex = NULL;
return false;
}
return true;
}
//==========================================================================
//
// D3DTex :: Update
//
// Copies image data from the underlying FTexture to the D3D texture.
//
//==========================================================================
bool D3DTex::Update()
{
D3DSURFACE_DESC desc;
D3DLOCKED_RECT lrect;
RECT rect;
assert(Tex != NULL);
assert(GameTex != NULL);
if (FAILED(Tex->GetLevelDesc(0, &desc)))
{
return false;
}
rect.left = 0;
rect.top = 0;
rect.right = GameTex->GetWidth();
rect.bottom = GameTex->GetHeight();
if (FAILED(Tex->LockRect(0, &lrect, &rect, 0)))
{
return false;
}
GameTex->FillBuffer((BYTE *)lrect.pBits, lrect.Pitch, rect.bottom, ToTexFmt(desc.Format));
Tex->UnlockRect(0);
return true;
}
//==========================================================================
//
// D3DTex :: GetTexFormat
//
// Returns the texture format that would best fit this texture.
//
//==========================================================================
D3DFORMAT D3DTex::GetTexFormat()
{
FTextureFormat fmt = GameTex->GetFormat();
IsGray = false;
switch (fmt)
{
case TEX_Pal: return D3DFMT_L8;
case TEX_Gray: IsGray = true; return D3DFMT_L8;
case TEX_RGB: return D3DFMT_A8R8G8B8;
case TEX_DXT1: return D3DFMT_DXT1;
case TEX_DXT2: return D3DFMT_DXT2;
case TEX_DXT3: return D3DFMT_DXT3;
case TEX_DXT4: return D3DFMT_DXT4;
case TEX_DXT5: return D3DFMT_DXT5;
default: I_FatalError ("GameTex->GetFormat() returned invalid format.");
}
return D3DFMT_L8;
}
//==========================================================================
//
// D3DTex :: ToTexFmt
//
// Converts a D3DFORMAT constant to something the FTexture system
// understands.
//
//==========================================================================
FTextureFormat D3DTex::ToTexFmt(D3DFORMAT fmt)
{
switch (fmt)
{
case D3DFMT_L8: return IsGray ? TEX_Gray : TEX_Pal;
case D3DFMT_A8R8G8B8: return TEX_RGB;
case D3DFMT_DXT1: return TEX_DXT1;
case D3DFMT_DXT2: return TEX_DXT2;
case D3DFMT_DXT3: return TEX_DXT3;
case D3DFMT_DXT4: return TEX_DXT4;
case D3DFMT_DXT5: return TEX_DXT5;
default:
assert(0); // LOL WUT?
return TEX_Pal;
}
}
- 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
//==========================================================================
//
// D3DPal Constructor
//
//==========================================================================
D3DPal::D3DPal(FRemapTable *remap, D3DFB *fb)
: Tex(NULL), Remap(remap)
{
int count;
// Attach to the palette list for the D3DFB
Next = fb->Palettes;
if (Next != NULL)
{
Next->Prev = &Next;
}
Prev = &fb->Palettes;
fb->Palettes = this;
- 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
// Palette textures must be 256 entries for Shader Model 1.4
if (fb->SM14)
{
count = 256;
}
else
{
int pow2count;
// Round up to the nearest power of 2.
for (pow2count = 1; pow2count < remap->NumEntries; pow2count <<= 1)
{ }
count = pow2count;
}
RoundedPaletteSize = count;
if (SUCCEEDED(fb->D3DDevice->CreateTexture(count, 1, 1, 0,
D3DFMT_A8R8G8B8, D3DPOOL_MANAGED, &Tex, NULL)))
{
if (!Update())
{
Tex->Release();
Tex = NULL;
}
}
}
//==========================================================================
//
// D3DPal Destructor
//
//==========================================================================
D3DPal::~D3DPal()
{
if (Tex != NULL)
{
Tex->Release();
Tex = NULL;
}
// Detach from the palette list
*Prev = Next;
if (Next != NULL)
{
Next->Prev = Prev;
}
// Remove link from the remap table
if (Remap != NULL)
{
Remap->Native = NULL;
}
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
}
//==========================================================================
//
// D3DPal :: Update
//
// Copies the palette to the texture.
//
//==========================================================================
bool D3DPal::Update()
{
D3DLOCKED_RECT lrect;
D3DCOLOR *buff;
const PalEntry *pal;
assert(Tex != NULL);
if (FAILED(Tex->LockRect(0, &lrect, NULL, 0)))
{
return false;
}
buff = (D3DCOLOR *)lrect.pBits;
pal = Remap->Palette;
// Should I allow the source palette to specify alpha values?
buff[0] = D3DCOLOR_ARGB(0, pal[0].r, pal[0].g, pal[0].b);
- 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
for (int i = 1; i < Remap->NumEntries; ++i)
{
buff[i] = D3DCOLOR_XRGB(pal[i].r, pal[i].g, pal[i].b);
- 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
}
Tex->UnlockRect(0);
return true;
}
//==========================================================================
//
// D3DFB :: Begin2D
//
// Begins 2D mode drawing operations. In particular, DrawTexture is
// rerouted to use Direct3D instead of the software renderer.
//
//==========================================================================
CVAR(Bool,test2d,true,0)
bool D3DFB::Begin2D()
{
if (!test2d) return false;
if (In2D)
{
return true;
}
In2D = 1;
Update();
In2D = 2;
return true;
}
- 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
//==========================================================================
//
// D3DFB :: CreateTexture
//
// Returns a native texture that wraps a FTexture.
//
//==========================================================================
FNativeTexture *D3DFB::CreateTexture(FTexture *gametex)
{
D3DTex *tex = new D3DTex(gametex, this);
- 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
if (tex->Tex == NULL)
{
delete tex;
return NULL;
}
return tex;
}
//==========================================================================
//
// D3DFB :: CreatePalette
//
// Returns a native texture that contains a palette.
//
//==========================================================================
FNativeTexture *D3DFB::CreatePalette(FRemapTable *remap)
{
D3DPal *tex = new D3DPal(remap, this);
if (tex->Tex == NULL)
{
delete tex;
return NULL;
}
return tex;
}
//==========================================================================
//
// D3DFB :: Clear
//
// Fills the specified region with a color.
//
//==========================================================================
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
void D3DFB::Clear (int left, int top, int right, int bottom, int palcolor, uint32 color)
{
if (In2D < 2)
{
Super::Clear(left, top, right, bottom, palcolor, color);
return;
}
if (palcolor >= 0)
{
color = GPalette.BaseColors[palcolor];
}
- 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
else if (APART(color) < 255)
{
Dim(color, APART(color)/255.f, left, top, right - left, bottom - top);
return;
}
D3DRECT rect = { left, top, right, bottom };
D3DDevice->Clear(1, &rect, D3DCLEAR_TARGET, color | 0xFF000000, 1.f, 0);
}
//==========================================================================
//
// D3DFB :: Dim
//
//==========================================================================
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
void D3DFB::Dim (PalEntry color, float amount, int x1, int y1, int w, int h)
{
if (amount <= 0)
return;
if (In2D < 2)
{
Super::Dim(color, amount, x1, y1, w, h);
return;
}
if (amount >= 1)
{
D3DRECT rect = { x1, y1, x1 + w, y1 + h };
D3DDevice->Clear(1, &rect, D3DCLEAR_TARGET, color | 0xFF000000, 1.f, 0);
}
else
{
FBVERTEX verts[4] =
{
{ x1-0.5f, y1-0.5f, 0.5f, 1, 0, 0 },
{ x1+w-0.5f, y1-0.5f, 0.5f, 1, 0, 0 },
{ x1+w-0.5f, y1+h-0.5f, 0.5f, 1, 0, 0 },
{ x1-0.5f, y1+h-0.5f, 0.5f, 1, 0, 0 }
};
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
SetAlphaBlend(TRUE, D3DBLEND_SRCALPHA, D3DBLEND_INVSRCALPHA);
SetPixelShader(DimShader);
SetConstant(1, color.r/255.f, color.g/255.f, color.b/255.f, amount);
D3DDevice->DrawPrimitiveUP(D3DPT_TRIANGLEFAN, 2, &verts, sizeof(FBVERTEX));
}
}
//==========================================================================
//
// D3DFB :: DrawTextureV
//
// If not in 2D mode, just call the normal software version.
// If in 2D mode, then use Direct3D calls to perform the drawing.
//
//==========================================================================
void STACK_ARGS D3DFB::DrawTextureV (FTexture *img, int x, int y, uint32 tags_first, va_list tags)
{
if (In2D < 2)
{
Super::DrawTextureV(img, x, y, tags_first, tags);
return;
}
DrawParms parms;
- 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
if (!ParseDrawTextureTags(img, x, y, tags_first, tags, &parms, true))
{
return;
}
D3DTex *tex = static_cast<D3DTex *>(img->GetNative());
if (tex == NULL)
{
assert(tex != NULL);
return;
}
float xscale = float(parms.destwidth) / parms.texwidth / 65536.f;
float yscale = float(parms.destheight) / parms.texheight / 65536.f;
float x0 = float(parms.x) / 65536.f - float(parms.left) * xscale;
float y0 = float(parms.y) / 65536.f - float(parms.top) * yscale;
float x1 = x0 + float(parms.destwidth) / 65536.f;
float y1 = y0 + float(parms.destheight) / 65536.f;
float u0 = 0.f;
float v0 = 0.f;
float u1 = 1.f;
float v1 = 1.f;
float uscale = 1.f / parms.texwidth / u1;
float vscale = 1.f / parms.texheight / v1 / yscale;
if (y0 < parms.uclip)
{
v0 += float(parms.uclip - y0) * vscale;
y0 = float(parms.uclip);
}
if (y1 > parms.dclip)
{
v1 -= float(y1 - parms.dclip) * vscale;
y1 = float(parms.dclip);
}
if (parms.flipX)
{
swap(u0, u1);
}
if (parms.windowleft > 0 || parms.windowright < parms.texwidth)
{
x0 += parms.windowleft * xscale;
u0 += parms.windowleft * uscale;
x1 -= (parms.texwidth - parms.windowright) * xscale;
u1 -= (parms.texwidth - parms.windowright) * uscale;
}
if (x0 < parms.lclip)
{
u0 += float(parms.lclip - x0) * uscale / xscale;
x0 = float(parms.lclip);
}
if (x1 > parms.rclip)
{
u1 -= float(x1 - parms.rclip) * uscale / xscale;
x1 = float(parms.rclip);
}
x0 -= 0.5f;
y0 -= 0.5f;
x1 -= 0.5f;
y1 -= 0.5f;
FBVERTEX verts[4] =
{
{ x0, y0, 0.5f, 1.f, u0, v0 },
{ x1, y0, 0.5f, 1.f, u1, v0 },
{ x1, y1, 0.5f, 1.f, u1, v1 },
{ x0, y1, 0.5f, 1.f, u0, v1 }
};
if (!SetStyle(tex, parms))
{
return;
}
- 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
SetTexture(0, tex->Tex);
D3DDevice->DrawPrimitiveUP(D3DPT_TRIANGLEFAN, 2, &verts, sizeof(FBVERTEX));
}
//==========================================================================
//
// D3DFB :: SetStyle
//
// Patterned after R_SetPatchStyle.
//
//==========================================================================
bool D3DFB::SetStyle(D3DTex *tex, DrawParms &parms)
{
D3DFORMAT fmt = tex->GetTexFormat();
- 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
ERenderStyle style = parms.style;
D3DBLEND fglevel, bglevel;
float alpha;
bool stencilling;
- 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
alpha = clamp<fixed_t> (parms.alpha, 0, FRACUNIT) / 65536.f;
if (style == STYLE_OptFuzzy)
{
style = STYLE_Translucent;
}
else if (style == STYLE_SoulTrans)
{
style = STYLE_Translucent;
alpha = transsouls;
}
// FIXME: STYLE_Fuzzy is not written
if (style == STYLE_Fuzzy)
{
style = STYLE_Translucent;
alpha = transsouls;
}
stencilling = false;
switch (style)
{
// Special modes
case STYLE_Shaded:
if (alpha > 0)
{
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
SetConstant(1, RPART(parms.fillcolor)/255.f, GPART(parms.fillcolor)/255.f, BPART(parms.fillcolor)/255.f, alpha);
SetPaletteTexture(ShadedPaletteTexture, 256);
SetAlphaBlend(TRUE, D3DBLEND_SRCALPHA, D3DBLEND_INVSRCALPHA);
return true;
}
return false;
// Standard modes
case STYLE_Stencil:
stencilling = true;
case STYLE_Normal:
fglevel = D3DBLEND_SRCALPHA;
bglevel = D3DBLEND_INVSRCALPHA;
alpha = 1;
break;
case STYLE_TranslucentStencil:
stencilling = true;
case STYLE_Translucent:
fglevel = D3DBLEND_SRCALPHA;
bglevel = D3DBLEND_INVSRCALPHA;
if (alpha == 0)
{
return false;
}
if (alpha == 1 && style == STYLE_Translucent)
{
style = STYLE_Normal;
}
break;
case STYLE_Add:
fglevel = D3DBLEND_SRCALPHA;
bglevel = D3DBLEND_ONE;
break;
default:
return false;
}
// Masking can only be turned off for STYLE_Normal, because it requires
// turning off the alpha blend.
- 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
if (!parms.masked && style == STYLE_Normal)
{
- 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
SetAlphaBlend(FALSE);
SetColorOverlay(parms.colorOverlay, 1);
if (fmt == D3DFMT_L8 && !tex->IsGray)
{
SetPaletteTexture(PaletteTexture, 256);
SetPixelShader(PalTexShader);
}
else
{
SetPixelShader(PlainShader);
}
}
else
{
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
SetAlphaBlend(TRUE, fglevel, bglevel);
if (!stencilling)
{
if (fmt == D3DFMT_L8)
- 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
{
if (parms.remap != NULL)
{
D3DPal *pal = reinterpret_cast<D3DPal *>(parms.remap->GetNative());
SetPaletteTexture(pal->Tex, pal->RoundedPaletteSize);
}
else if (tex->IsGray)
{
SetPixelShader(PlainShader);
}
else
{
SetPaletteTexture(PaletteTexture, 256);
}
SetPixelShader(PalTexShader);
- 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
}
else
{
SetPixelShader(PlainShader);
- 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
}
SetColorOverlay(parms.colorOverlay, alpha);
}
else
{
SetConstant(1,
RPART(parms.fillcolor)/255.f,
GPART(parms.fillcolor)/255.f,
BPART(parms.fillcolor)/255.f, alpha);
if (fmt == D3DFMT_L8)
{
SetPaletteTexture(StencilPaletteTexture, 256);
SetPixelShader(PalTexShader);
}
else
{
SetPixelShader(PlainStencilShader);
}
}
}
return true;
- Discovered that Shader Model 1.4 clamps my constants, so I can't use palettes smaller than 256 entries with the shader I wrote for it. Is there a list of gotchas like this listed some where? I'd really like to see it. Well, when compiled with SM2.0, the PalTex shader seems to be every-so- slightly faster on my GF7950GT than the SM1.4 version, so I guess it's a minor win for cards that support it. - Fixed: ST_Endoom() failed to free the bitmap it used. - Added the DTA_ColorOverlay attribute to blend a color with the texture being drawn. For software, this (currently) only works with black. For hardware, it works with any color. The motiviation for this was so I could rewrite the status bar calls that passed DIM_MAP to DTA_Translation to draw darker icons into something that didn't require making a whole new remap table. - After having an "OMG! How could I have been so stupid?" moment, I have removed the off-by-one check from D3DFB. I had thought the off-by-one error was caused by rounding errors by the shader hardware. Not so. Rather, I wasn't sampling what I thought I was sampling. A texture that uses palette index 255 passes the value 1.0 to the shader. The shader needs to adjust the range of its palette indexes, or it will end up trying to read color 256 from the palette texture when it should be reading color 255. Doh! - The TranslationToTable() function has been added to map from translation numbers used by actors to the tables those numbers represent. This function performs validation for the input and returns NULL if the input value is invalid. - Major changes to the way translation tables work: No longer are they each a 256-byte array. Instead, the FRemapTable structure is used to represent each one. It includes a remap array for the software renderer, a palette array for a hardware renderer, and a native texture pointer for D3DFB. The translationtables array itself is now an array of TArrays that point to the real tables. The DTA_Translation attribute must also be passed a pointer to a FRemapTable, not a byte array as previously. - Modified DFrameBuffer::DrawRateStuff() so that it can do its thing properly for D3DFB's 2D mode. Before, any fullscreen graphics (like help images) covered it up. SVN r640 (trunk)
2007-12-26 04:42:15 +00:00
}
void D3DFB::SetColorOverlay(DWORD color, float alpha)
{
if (APART(color) != 0)
{
float a = 255.f / APART(color);
float r = RPART(color) * a;
float g = GPART(color) * a;
float b = BPART(color) * a;
SetConstant(0, r, g, b, 0);
a = 1 - 1 / a;
SetConstant(1, a, a, a, alpha);
}
else
{
SetConstant(0, 0, 0, 0, 0);
SetConstant(1, 1, 1, 1, alpha);
}
}
void D3DFB::SetAlphaBlend(BOOL enabled, D3DBLEND srcblend, D3DBLEND destblend)
{
if (!enabled)
{ // Disable alpha blend
if (AlphaBlendEnabled)
{
AlphaBlendEnabled = FALSE;
D3DDevice->SetRenderState(D3DRS_ALPHABLENDENABLE, FALSE);
}
}
else
{ // Enable alpha blend
assert(srcblend != 0);
assert(destblend != 0);
if (!AlphaBlendEnabled)
{
AlphaBlendEnabled = TRUE;
D3DDevice->SetRenderState(D3DRS_ALPHABLENDENABLE, TRUE);
}
if (AlphaSrcBlend != srcblend)
{
AlphaSrcBlend = srcblend;
D3DDevice->SetRenderState(D3DRS_SRCBLEND, srcblend);
}
if (AlphaDestBlend != destblend)
{
AlphaDestBlend = destblend;
D3DDevice->SetRenderState(D3DRS_DESTBLEND, destblend);
}
}
}
void D3DFB::SetConstant(int cnum, float r, float g, float b, float a)
{
if (Constant[cnum][0] != r ||
Constant[cnum][1] != g ||
Constant[cnum][2] != b ||
Constant[cnum][3] != a)
{
Constant[cnum][0] = r;
Constant[cnum][1] = g;
Constant[cnum][2] = b;
Constant[cnum][3] = a;
D3DDevice->SetPixelShaderConstantF(cnum, Constant[cnum], 1);
}
}
void D3DFB::SetPixelShader(IDirect3DPixelShader9 *shader)
{
if (CurPixelShader != shader)
{
CurPixelShader = shader;
D3DDevice->SetPixelShader(shader);
}
}
void D3DFB::SetTexture(int tnum, IDirect3DTexture9 *texture)
{
if (Texture[tnum] != texture)
{
Texture[tnum] = texture;
D3DDevice->SetTexture(tnum, texture);
}
}
void D3DFB::SetPaletteTexture(IDirect3DTexture9 *texture, int count)
{
if (count == 256 || SM14)
- 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
{
// Shader Model 1.4 only uses 256-color palettes.
SetConstant(2, 255 / 256.f, 0.5f / 256.f, 0, 0);
}
else
{
// The pixel shader receives color indexes in the range [0.0,1.0].
// The palette texture is also addressed in the range [0.0,1.0],
// HOWEVER the coordinate 1.0 is the right edge of the texture and
// not actually the texture itself. We need to scale and shift
// the palette indexes so they lie exactly in the center of each
// texel. For a normal palette with 256 entries, that means the
// range we use should be [0.5,255.5], adjusted so the coordinate
// is still with [0.0,1.0].
//
// The constant register c2 is used to hold the multiplier in the
// x part and the adder in the y part.
float fcount = 1 / float(count);
SetConstant(2, 255 * fcount, 0.5f * fcount, 0, 0);
}
SetTexture(1, texture);
SetPixelShader(PalTexShader);
}