libretro/libretro-tyrquake
1
0
Fork 0
mirror of https://github.com/libretro/libretro-tyrquake.git synced 2024-11-27 10:10:58 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
af3aa95712
libretro-tyrquake/common/vid_sdl.c

678 lines
16 KiB
C
Raw Normal View History

video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
/*
Copyright (C) 1996-1997 Id Software, Inc.
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <stdlib.h>
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#include "SDL.h"
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
#include "cdaudio.h"
#include "cmd.h"
#include "common.h"
#include "console.h"
#include "cvar.h"
#include "d_iface.h"
#include "d_local.h"
#include "draw.h"
#include "input.h"
#include "keys.h"
#include "menu.h"
#include "quakedef.h"
#include "screen.h"
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#include "sdl_common.h"
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
#include "sound.h"
#include "sys.h"
#include "vid.h"
#include "view.h"
#include "wad.h"
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifdef _WIN32
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
#include "winquake.h"
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#endif
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
#ifdef NQ_HACK
#include "host.h"
#endif
#ifdef QW_HACK
#include "client.h"
#endif
// FIXME: evil hack to get full DirectSound support with SDL
#ifdef _WIN32
#include <windows.h>
HWND mainwindow;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
qboolean DDActive = false;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
#endif
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
static SDL_Renderer *renderer = NULL;
static SDL_Texture *texture = NULL;
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
static SDL_PixelFormat *sdl_format = NULL;
vid: use desktop pixel format for windowed modes Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:44:36 +00:00
static SDL_PixelFormat *sdl_desktop_format = NULL;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
/* ------------------------------------------------------------------------- */
static byte *vid_surfcache;
static int vid_surfcachesize;
static int VID_highhunkmark;
static int window_width, window_height;
unsigned short d_8to16table[256];
unsigned d_8to24table[256];
viddef_t vid; /* global video state */
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifdef _WIN32
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
static HICON hIcon;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
RECT window_rect;
#endif
vid: implement basic VID_Shutdown for sdl Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:45:12 +00:00
void VID_Shutdown(void)
{
if (renderer)
SDL_DestroyRenderer(renderer);
if (sdl_window)
SDL_DestroyWindow(sdl_window);
if (sdl_format)
SDL_FreeFormat(sdl_format);
if (sdl_desktop_format)
SDL_FreeFormat(sdl_desktop_format);
}
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
static qboolean palette_changed;
void
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
VID_ShiftPalette(const byte *palette)
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
{
VID_SetPalette(palette);
}
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
void VID_SetDefaultMode(void) { }
static cvar_t vid_mode = {
.name = "vid_mode",
.string = stringify(VID_MODE_WINDOWED),
.archive = false
};
static cvar_t _vid_default_mode = {
.name = "_vid_default_mode",
.string = stringify(VID_MODE_WINDOWED),
.archive = true
};
static cvar_t _vid_default_mode_win = {
.name = "_vid_default_mode_win",
.string = stringify(VID_MODE_FULLSCREEN_DEFAULT),
.archive = true
};
static cvar_t vid_fullscreen_mode = {
.name = "vid_fullscreen_mode",
.string = stringify(VID_MODE_FULLSCREEN_DEFAULT),
.archive = true
};
static cvar_t vid_windowed_mode = {
.name = "vid_windowed_mode",
.string = stringify(VID_MODE_WINDOWED),
.archive = true
};
static cvar_t vid_wait = { "vid_wait", "0" };
static cvar_t vid_nopageflip = { "vid_nopageflip", "0", true };
static cvar_t _vid_wait_override = { "_vid_wait_override", "0", true };
static cvar_t block_switch = { "block_switch", "0", true };
static cvar_t vid_window_x = { "vid_window_x", "0", true };
static cvar_t vid_window_y = { "vid_window_y", "0", true };
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
int vid_modenum = VID_MODE_NONE;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
typedef struct {
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
typeof(SDL_PIXELFORMAT_UNKNOWN) format;
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
} qvidformat_t;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
static qboolean Minimized;
qboolean
window_visible(void)
{
return !Minimized;
}
static void
VID_InitModeList(void)
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
int i, err;
int displays, sdlmodes;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
SDL_DisplayMode sdlmode;
qvidmode_t *mode;
qvidformat_t *format;
vid: quick hack for SDL command line vid mode settings This could be done so much better, but for now just want a quick hack so that the launcher app can pass in vid mode settings and the thing will actually do something useful! Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-27 06:57:07 +00:00
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
displays = SDL_GetNumVideoDisplays();
if (displays < 1)
Sys_Error("%s: no displays found (%s)", __func__, SDL_GetError());
/* FIXME - allow use of more than one display */
vid: sdl display index is zero-based Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-08 02:11:43 +00:00
sdlmodes = SDL_GetNumDisplayModes(0);
if (sdlmodes < 0)
Con_SafePrintf("%s: error enumerating SDL display modes (%s)\n",
__func__, SDL_GetError());
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: set up windowed mode list for sdl Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-19 23:30:19 +00:00
/*
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
* Check availability of fullscreen modes
vid: sdl display index is zero-based Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-08 02:11:43 +00:00
* (default to display 0 for now)
vid: set up windowed mode list for sdl Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-19 23:30:19 +00:00
*/
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
mode = &modelist[1];
nummodes = 1;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
for (i = 0; i < sdlmodes && nummodes < MAX_MODE_LIST; i++) {
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
err = SDL_GetDisplayMode(0, i, &sdlmode);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
if (err)
Sys_Error("%s: couldn't get mode %d info (%s)",
__func__, i, SDL_GetError());
printf("%s: checking mode %i: %dx%d, %s\n", __func__,
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
i, sdlmode.w, sdlmode.h, SDL_GetPixelFormatName(sdlmode.format));
vid: set up windowed mode list for sdl Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-19 23:30:19 +00:00
vid: correctly set software MAXWIDTH The math used for edge sorting overflows at width 2048 due to use of a 12.20 fixed-point numeric type. Also, make sure we check MAXWIDTH in the vid_sdl driver. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-08 00:47:26 +00:00
if (sdlmode.h > MAXHEIGHT || sdlmode.w > MAXWIDTH)
vid: accept any 8888 sdl packed pixel layout Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-08 02:12:37 +00:00
continue;
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
if (SDL_PIXELTYPE(sdlmode.format) == SDL_PIXELTYPE_PACKED32)
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
modelist[nummodes].bpp = 32;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
else if (SDL_PIXELTYPE(sdlmode.format) == SDL_PIXELTYPE_PACKED16)
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
modelist[nummodes].bpp = 16;
else
vid: respect MAXHEIGHT in vid mode selection Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-20 06:35:11 +00:00
continue;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
mode->modenum = nummodes;
mode->width = sdlmode.w;
mode->height = sdlmode.h;
mode->refresh = sdlmode.refresh_rate;
format = (qvidformat_t *)mode->driverdata;
format->format = sdlmode.format;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
nummodes++;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
mode++;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
vid: quick hack for SDL command line vid mode settings This could be done so much better, but for now just want a quick hack so that the launcher app can pass in vid mode settings and the thing will actually do something useful! Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-27 06:57:07 +00:00
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
VID_SortModeList(modelist, nummodes);
vid: quick hack for SDL command line vid mode settings This could be done so much better, but for now just want a quick hack so that the launcher app can pass in vid mode settings and the thing will actually do something useful! Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-27 06:57:07 +00:00
}
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
/*
====================
VID_CheckAdequateMem
====================
*/
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
qboolean
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
VID_CheckAdequateMem(int width, int height)
{
int tbuffersize;
tbuffersize = width * height * sizeof(*d_pzbuffer);
tbuffersize += D_SurfaceCacheForRes(width, height);
/*
* see if there's enough memory, allowing for the normal mode 0x13 pixel,
* z, and surface buffers
*/
if ((host_parms.memsize - tbuffersize + SURFCACHE_SIZE_AT_320X200 +
0x10000 * 3) < minimum_memory)
return false;
return true;
}
/*
================
VID_AllocBuffers
================
*/
static qboolean
VID_AllocBuffers(int width, int height)
{
int tsize, tbuffersize;
tsize = D_SurfaceCacheForRes(width, height);
tbuffersize = width * height * sizeof(*d_pzbuffer);
tbuffersize += tsize;
/*
* see if there's enough memory, allowing for the normal mode 0x13 pixel,
* z, and surface buffers
*/
if ((host_parms.memsize - tbuffersize + SURFCACHE_SIZE_AT_320X200 +
0x10000 * 3) < minimum_memory) {
Con_SafePrintf("Not enough memory for video mode\n");
return false;
}
vid_surfcachesize = tsize;
if (d_pzbuffer) {
D_FlushCaches();
Hunk_FreeToHighMark(VID_highhunkmark);
d_pzbuffer = NULL;
}
VID_highhunkmark = Hunk_HighMark();
d_pzbuffer = Hunk_HighAllocName(tbuffersize, "video");
vid_surfcache = (byte *)d_pzbuffer + width * height * sizeof(*d_pzbuffer);
return true;
}
qboolean
VID_IsFullScreen()
{
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
return !!vid_modenum;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
/*
================
VID_UpdateWindowStatus
================
*/
static void
VID_UpdateWindowStatus(void)
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifdef _WIN32
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
window_rect.left = window_x;
window_rect.top = window_y;
window_rect.right = window_x + window_width;
window_rect.bottom = window_y + window_height;
window_center_x = (window_rect.left + window_rect.right) / 2;
window_center_y = (window_rect.top + window_rect.bottom) / 2;
IN_UpdateClipCursor();
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#endif
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
qboolean
VID_SetMode(const qvidmode_t *mode, const byte *palette)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
{
vid: Make fullscreen mode switching work too Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-20 06:25:51 +00:00
Uint32 flags;
input: fix sdl mouse input Tried to follow similar patterns to those in the X11 input driver. Needed an ugly hack to re-acquire mouse grabs on mode change, since the APIs aren't set up properly to conditionally release and re-grab... yet. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 06:54:12 +00:00
qboolean mouse_grab;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
const qvidformat_t *format;
input: fix sdl mouse input Tried to follow similar patterns to those in the X11 input driver. Needed an ugly hack to re-acquire mouse grabs on mode change, since the APIs aren't set up properly to conditionally release and re-grab... yet. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 06:54:12 +00:00
/* FIXME - hack to reset mouse grabs */
mouse_grab = _windowed_mouse.value;
if (mouse_grab) {
_windowed_mouse.value = 0;
_windowed_mouse.callback(&_windowed_mouse);
}
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
flags = SDL_WINDOW_SHOWN;
vid: be more consistent with fullscreen mode test Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-02 02:49:16 +00:00
if (mode != modelist)
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
flags |= SDL_WINDOW_FULLSCREEN;
if (renderer)
SDL_DestroyRenderer(renderer);
if (sdl_window)
SDL_DestroyWindow(sdl_window);
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
if (sdl_format)
SDL_FreeFormat(sdl_format);
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
format = (const qvidformat_t *)mode->driverdata;
sdl_format = SDL_AllocFormat(format->format);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
sdl_window = SDL_CreateWindow("TyrQuake",
SDL_WINDOWPOS_UNDEFINED,
SDL_WINDOWPOS_UNDEFINED,
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
mode->width, mode->height, flags);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
if (!sdl_window)
Sys_Error("%s: Unable to create window: %s", __func__, SDL_GetError());
renderer = SDL_CreateRenderer(sdl_window, -1, SDL_RENDERER_ACCELERATED);
if (!renderer)
Sys_Error("%s: Unable to create renderer: %s", __func__, SDL_GetError());
texture = SDL_CreateTexture(renderer,
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
format->format,
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
SDL_TEXTUREACCESS_STREAMING,
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
mode->width, mode->height);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
if (!texture)
Sys_Error("%s: Unable to create texture: %s", __func__, SDL_GetError());
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
//VID_InitGamma(palette);
VID_SetPalette(palette);
vid.numpages = 1;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
vid.width = vid.conwidth = mode->width;
vid.height = vid.conheight = mode->height;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid.maxwarpwidth = WARP_WIDTH;
vid.maxwarpheight = WARP_HEIGHT;
vid.aspect = ((float)vid.height / (float)vid.width) * (320.0 / 240.0);
vid.colormap = host_colormap;
vid.fullbright = 256 - LittleLong(*((int *)vid.colormap + 2048));
VID_AllocBuffers(vid.width, vid.height);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
// In-memory buffer which we upload via SDL texture
vid.buffer = vid.conbuffer = vid.direct = Hunk_HighAllocName(vid.width * vid.height, "vidbuf");
vid.rowbytes = vid.conrowbytes = vid.width;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
D_InitCaches(vid_surfcache, vid_surfcachesize);
window_width = vid.width;
window_height = vid.height;
VID_UpdateWindowStatus();
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
vid_modenum = mode - modelist;
Cvar_SetValue("vid_mode", vid_modenum);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid.recalc_refdef = 1;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifdef _WIN32
vid: Implement (windowed) video mode changes via menu Video mode changes via the menu now work (at least for windowed modes). Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-20 05:55:42 +00:00
mainwindow=GetActiveWindow();
SendMessage(mainwindow, WM_SETICON, (WPARAM)ICON_BIG, (LPARAM)hIcon);
SendMessage(mainwindow, WM_SETICON, (WPARAM)ICON_SMALL, (LPARAM)hIcon);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#endif
vid: Implement (windowed) video mode changes via menu Video mode changes via the menu now work (at least for windowed modes). Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-20 05:55:42 +00:00
input: fix sdl mouse input Tried to follow similar patterns to those in the X11 input driver. Needed an ugly hack to re-acquire mouse grabs on mode change, since the APIs aren't set up properly to conditionally release and re-grab... yet. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 06:54:12 +00:00
/* FIXME - hack to reset mouse grabs */
if (mouse_grab) {
_windowed_mouse.value = 1;
_windowed_mouse.callback(&_windowed_mouse);
}
vid: be more consistent with fullscreen mode test Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-02 02:49:16 +00:00
Cvar_SetValue("vid_fullscreen", mode != modelist);
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
Cvar_SetValue("vid_width", mode->width);
Cvar_SetValue("vid_height", mode->height);
Cvar_SetValue("vid_bpp", mode->bpp);
Cvar_SetValue("vid_refreshrate", mode->refresh);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
return true;
}
/* ------------------------------------------------------------------------- */
// The original defaults
#define BASEWIDTH 320
#define BASEHEIGHT 200
byte *VGA_pagebase;
int VGA_width, VGA_height, VGA_rowbytes, VGA_bufferrowbytes = 0;
void
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
VID_SetPalette(const byte *palette)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
unsigned i, r, g, b;
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
switch (SDL_PIXELTYPE(sdl_format->format)) {
case SDL_PIXELTYPE_PACKED32:
for (i = 0; i < 256; i++) {
r = palette[0];
g = palette[1];
b = palette[2];
palette += 3;
d_8to24table[i] = SDL_MapRGB(sdl_format, r, g, b);
}
break;
case SDL_PIXELTYPE_PACKED16:
for (i = 0; i < 256; i++) {
r = palette[0];
g = palette[1];
b = palette[2];
palette += 3;
d_8to16table[i] = SDL_MapRGB(sdl_format, r, g, b);
}
break;
default:
Sys_Error("%s: unsupported pixel format (%s)", __func__,
SDL_GetPixelFormatName(sdl_format->format));
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
palette_changed = true;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
void
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
VID_Init(const byte *palette)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
{
vid: use desktop pixel format for windowed modes Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:44:36 +00:00
int err;
SDL_DisplayMode desktop_mode;
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
qvidmode_t *mode;
qvidformat_t *format;
const qvidmode_t *setmode;
vid: use desktop pixel format for windowed modes Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:44:36 +00:00
vid: init cvars before first mode set Otherwise, the vid_mode cvar gets inited to zero and we change immediately back to mode 0. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-27 06:52:59 +00:00
Cvar_RegisterVariable(&vid_mode);
Cvar_RegisterVariable(&vid_wait);
Cvar_RegisterVariable(&vid_nopageflip);
Cvar_RegisterVariable(&_vid_wait_override);
Cvar_RegisterVariable(&_vid_default_mode);
Cvar_RegisterVariable(&_vid_default_mode_win);
Cvar_RegisterVariable(&vid_fullscreen_mode);
Cvar_RegisterVariable(&vid_windowed_mode);
Cvar_RegisterVariable(&block_switch);
Cvar_RegisterVariable(&vid_window_x);
Cvar_RegisterVariable(&vid_window_y);
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
VID_InitModeCvars();
vid: init cvars before first mode set Otherwise, the vid_mode cvar gets inited to zero and we change immediately back to mode 0. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-27 06:52:59 +00:00
Cmd_AddCommand("vid_describemodes", VID_DescribeModes_f);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
/*
* Init SDL and the video subsystem
*/
Q_SDL_InitOnce();
vid: get the basic sgl video target up and running Hard coded to just one resolution with no mode changes, but works pretty well (tested only on Mac) within those limitations. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-11 05:54:26 +00:00
err = SDL_InitSubSystem(SDL_INIT_VIDEO);
if (err < 0)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
Sys_Error("VID: Couldn't load SDL: %s", SDL_GetError());
vid: use desktop pixel format for windowed modes Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:44:36 +00:00
err = SDL_GetDesktopDisplayMode(0, &desktop_mode);
if (err)
Sys_Error("%s: Unable to query desktop display mode (%s)",
__func__, SDL_GetError());
sdl_desktop_format = SDL_AllocFormat(desktop_mode.format);
if (!sdl_desktop_format)
Sys_Error("%s: Unable to allocate desktop pixel format (%s)",
__func__, SDL_GetError());
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
/* Init default windowed mode */
mode = modelist;
mode->modenum = 0;
mode->bpp = sdl_desktop_format->BitsPerPixel;
format = (qvidformat_t *)mode->driverdata;
format->format = sdl_desktop_format->format;
mode->refresh = desktop_mode.refresh_rate;
mode->width = 640;
mode->height = 480;
nummodes = 1;
/* TODO: read config files first to avoid multiple mode sets */
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
VID_InitModeList();
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
setmode = VID_GetCmdlineMode();
if (!setmode)
setmode = &modelist[0];
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: implement replacement video options menu Replace the current video mode menu with one modelled after the FitzQuake video options menu. This changeset is far larger than I usually would like, but I'll try to summarise the main points below: * The recent hack for command line mode setting got replaced * The video driver now dumps all modes into modelist and calls a sort function when completed. Modes are sorted by height, width, bpp and refresh rate (in that order). * vid mode 0 is now the only windowed mode and we update the width/height when a different resolution is requested. * vmode_t replaced with qvidmode_t and re-thought. Drivers can stuff a small amount of data, currently only used by the SDL driver to save the pixel format for each mode. * Some old command line options removed: -allow360, -noforcevga, -noadjustaspect. While everything currently compiles, unfortunately some features are currently broken. Here is a bit of a TODO list: * Mode testing is currently broken (was broken on SDL previously) * Command line mode setting is inconsistent and somewhat broken * Sound gets messed up on vid_win.c and vid_wgl.c. The workaround in vid_win.c seems effective there, but not in wgl, so it's probably not the right fix in the end. * x11 drivers only have compatiblity glue around the new menus Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-04-01 03:13:49 +00:00
VID_SetMode(setmode, palette);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifdef _WIN32
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
mainwindow=GetActiveWindow();
SendMessage(mainwindow, WM_SETICON, (WPARAM)ICON_BIG, (LPARAM)hIcon);
SendMessage(mainwindow, WM_SETICON, (WPARAM)ICON_SMALL, (LPARAM)hIcon);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#endif
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid_menudrawfn = VID_MenuDraw;
vid_menukeyfn = VID_MenuKey;
}
void
VID_Update(vrect_t *rects)
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
SDL_Rect subrect;
int i;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vrect_t *rect;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
vrect_t fullrect;
byte *src;
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
void *dst;
Uint32 *dst32;
Uint16 *dst16;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
int pitch;
int height;
int err;
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
const qvidmode_t *mode;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: check for vid_mode changes in sdl target Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-29 10:42:19 +00:00
/*
* Check for vid_mode changes
*/
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
if ((int)vid_mode.value != vid_modenum) {
vid: start factoring out video mode handling from drivers Separate some of the common code out into vid_mode.c. The interfaces are still a bit clunky, but now at least sharing the video menu code between the Windows and SDL drivers. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2013-03-28 05:08:42 +00:00
mode = &modelist[(int)vid_mode.value];
VID_SetMode(mode, host_basepal);
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
/* FIXME - not the right place! redraw the scene to buffer first */
return;
}
vid: check for vid_mode changes in sdl target Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-29 10:42:19 +00:00
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
/*
* If the palette changed, refresh the whole screen
*/
if (palette_changed) {
palette_changed = false;
fullrect.x = 0;
fullrect.y = 0;
fullrect.width = vid.width;
fullrect.height = vid.height;
fullrect.pnext = NULL;
rects = &fullrect;
}
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
for (rect = rects; rect; rect = rect->pnext) {
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
subrect.x = rect->x;
subrect.y = rect->y;
subrect.w = rect->width;
subrect.h = rect->height;
err = SDL_LockTexture(texture, &subrect, (void **)&dst, &pitch);
if (err)
Sys_Error("%s: unable to lock texture (%s)",
__func__, SDL_GetError());
src = vid.buffer + rect->y * vid.width + rect->x;
height = subrect.h;
vid: enable sdl to render to 16 or 32 bit formats Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-09 02:43:43 +00:00
switch (SDL_PIXELTYPE(sdl_format->format)) {
case SDL_PIXELTYPE_PACKED32:
dst32 = dst;
while (height--) {
for (i = 0; i < rect->width; i++)
dst32[i] = d_8to24table[src[i]];
dst32 += pitch / sizeof(*dst32);
src += vid.width;
}
break;
case SDL_PIXELTYPE_PACKED16:
dst16 = dst;
while (height--) {
for (i = 0; i < rect->width; i++)
dst16[i] = d_8to16table[src[i]];
dst16 += pitch / sizeof(*dst16);
src += vid.width;
}
break;
default:
Sys_Error("%s: unsupported pixel format (%s)", __func__,
SDL_GetPixelFormatName(sdl_format->format));
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
}
SDL_UnlockTexture(texture);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
err = SDL_RenderCopy(renderer, texture, NULL, NULL);
if (err)
Sys_Error("%s: unable to render texture (%s)", __func__, SDL_GetError());
SDL_RenderPresent(renderer);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
void
vid: constify source bitmap for D_BeginDirectRect Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-22 05:44:59 +00:00
D_BeginDirectRect(int x, int y, const byte *pbitmap, int width, int height)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
int err, i;
vid: constify source bitmap for D_BeginDirectRect Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-22 05:44:59 +00:00
const byte *src;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
unsigned *dst;
int pitch;
SDL_Rect subrect;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
if (!texture || !renderer)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
return;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
subrect.x = (x < 0) ? vid.width + x - 1 : x;
subrect.y = y;
subrect.w = width;
subrect.h = height;
err = SDL_LockTexture(texture, &subrect, (void **)&dst, &pitch);
if (err)
Sys_Error("%s: unable to lock texture (%s)", __func__, SDL_GetError());
src = pbitmap;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
while (height--) {
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
for (i = 0; i < width; i++)
dst[i] = d_8to24table[src[i]];
dst += pitch / sizeof(*dst);
src += width;
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
SDL_UnlockTexture(texture);
err = SDL_RenderCopy(renderer, texture, NULL, NULL);
if (err)
Sys_Error("%s: unable to render texture (%s)", __func__, SDL_GetError());
SDL_RenderPresent(renderer);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
void
D_EndDirectRect(int x, int y, int width, int height)
{
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
int err, i;
byte *src;
unsigned *dst;
int pitch;
SDL_Rect subrect;
if (!texture || !renderer)
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
return;
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
subrect.x = (x < 0) ? vid.width + x - 1 : x;
subrect.y = y;
subrect.w = width;
subrect.h = height;
err = SDL_LockTexture(texture, &subrect, (void **)&dst, &pitch);
if (err)
Sys_Error("%s: unable to lock texture (%s)", __func__, SDL_GetError());
src = vid.buffer + y * vid.width + subrect.x;
while (height--) {
for (i = 0; i < width; i++)
dst[i] = d_8to24table[src[i]];
dst += pitch / sizeof(*dst);
src += vid.width;
}
SDL_UnlockTexture(texture);
err = SDL_RenderCopy(renderer, texture, NULL, NULL);
if (err)
Sys_Error("%s: unable to render texture (%s)", __func__, SDL_GetError());
SDL_RenderPresent(renderer);
video: Add experimental SDL renderer for Win32 This is a rather large patch which adds an experimental SDL rendering target for Win32 (software renderer only). Changes to support SDL input spilled over into a number of other areas - most notably the big churn in keys.h to match the key symbols to those of SDL. Credit to the Quakeforge project, whose sources were very helpful for me in working out how this all fits together. Compile with USE_SDL=Y to select the experimental target (tyr-quake.exe and tyr-qwcl.exe only). Expect that not everything works - e.g. A single 800x600 video mode is hard coded right now. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-10-07 07:09:29 +00:00
}
void
VID_LockBuffer(void)
{
}
void
VID_UnlockBuffer(void)
{
}
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
#ifndef _WIN32
void
Sys_SendKeyEvents(void)
{
input: fix hang on new game screen for sdl/sgl on other platforms Fixing the same bug in multiple places... bit of a hint that some refactoring is in order here. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-16 10:28:55 +00:00
IN_ProcessEvents();
vid: update the sdl driver(s) to the SDL2 API Still quite a few things broken, especially mouse input, but video works much better and now compiles and works on OS X too. We now expect sdl2-config in the path and use that in the Makefile instead of the old SDLBASE guesswork. I broke out a couple of things into sdl_common.[ch] just to coordinate init and share the main window pointer. Still quite some limitations in the driver, only supports RBG888 modes right now, but adding RGB565 should be straight forward. Still quite a bit slower than the Win/MGL driver... would like to understand why that is at some point. But it's already much faster than the SDL1.2 version. Signed-off-by: Kevin Shanahan <kmshanah@disenchant.net>
2012-12-07 21:04:49 +00:00
}
#endif
Reference in New Issue Copy Permalink