qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-26 22:31:05 +00:00
Code Issues Projects Releases Packages Wiki Activity
quakeforge/libs/video/targets/context_win.c

1171 lines
26 KiB
C
Raw Normal View History

Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
/*
vid_win.c
Win32 vid component
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:
Free Software Foundation, Inc.
59 Temple Place - Suite 330
Boston, MA 02111-1307, USA
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#include "winquake.h"
#include <ddraw.h>
#include "QF/cmd.h"
#include "QF/cvar.h"
#include "QF/input.h"
#include "QF/qargs.h"
#include "QF/sys.h"
#include "QF/va.h"
#include "QF/vid.h"
#include "context_win.h"
#include "r_shared.h"
#include "vid_internal.h"
#include "vid_sw.h"
HWND win_mainwindow;
HDC win_maindc;
int win_palettized;
[win] Clean up a pile of messy externs
2021-03-31 01:36:30 +00:00
int win_minimized;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
int win_canalttab = 0;
sw_ctx_t *win_sw_context;
#define MODE_WINDOWED 0
#define MODE_SETTABLE_WINDOW 2
#define NO_MODE (MODE_WINDOWED - 1)
#define MODE_FULLSCREEN_DEFAULT (MODE_WINDOWED + 3)
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
int vid_ddraw;
static cvar_t vid_ddraw_cvar = {
.name = "vid_ddraw",
.description =
"",
.default_value = "1",
.flags = CVAR_NONE,
.value = { .type = &cexpr_int, .value = &vid_ddraw },
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
// Note that 0 is MODE_WINDOWED
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static int vid_mode;
static cvar_t vid_mode_cvar = {
.name = "vid_mode",
.description =
"",
.default_value = "0",
.flags = CVAR_NONE,
.value = { .type = &cexpr_int, .value = &vid_mode },
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
// Note that 0 is MODE_WINDOWED
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int _vid_default_mode;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t _vid_default_mode_cvar = {
.name = "_vid_default_mode",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &_vid_default_mode },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
// Note that 3 is MODE_FULLSCREEN_DEFAULT
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static int _vid_default_mode_win;
static cvar_t _vid_default_mode_win_cvar = {
.name = "_vid_default_mode_win",
.description =
"",
.default_value = "3",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &_vid_default_mode_win },
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int vid_wait;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t vid_wait_cvar = {
.name = "vid_wait",
.description =
"",
.default_value = "0",
.flags = CVAR_NONE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &vid_wait },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int vid_nopageflip;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t vid_nopageflip_cvar = {
.name = "vid_nopageflip",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &vid_nopageflip },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int _vid_wait_override;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t _vid_wait_override_cvar = {
.name = "_vid_wait_override",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &_vid_wait_override },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int vid_config_x;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t vid_config_x_cvar = {
.name = "vid_config_x",
.description =
"",
.default_value = "800",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &vid_config_x },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int vid_config_y;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t vid_config_y_cvar = {
.name = "vid_config_y",
.description =
"",
.default_value = "600",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &vid_config_y },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int vid_stretch_by_2;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t vid_stretch_by_2_cvar = {
.name = "vid_stretch_by_2",
.description =
"",
.default_value = "1",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &vid_stretch_by_2 },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
static int _windowed_mouse;
static cvar_t _windowed_mouse_cvar = {
.name = "_windowed_mouse",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &_windowed_mouse },
};
static int vid_fullscreen_mode;
static cvar_t vid_fullscreen_mode_cvar = {
.name = "vid_fullscreen_mode",
.description =
"",
.default_value = "3",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &vid_fullscreen_mode },
};
static int vid_windowed_mode;
static cvar_t vid_windowed_mode_cvar = {
.name = "vid_windowed_mode",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &vid_windowed_mode },
};
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
int block_switch;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
static cvar_t block_switch_cvar = {
.name = "block_switch",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
[cvar] Clean up most misinterpreted cvar types The misinterpretations were due to either the cvar not being accessed directly by the engine, but via only the callback, or the cvars were accesssed only by progs (in which case, they should be float). The remainder are a potential enum (hud gravity) and a "too hard basket" (rcon password: need to figure out how I want to handle secret strings).
2022-04-24 11:04:06 +00:00
.value = { .type = &cexpr_int, .value = &block_switch },
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
};
static int vid_window_x;
static cvar_t vid_window_x_cvar = {
.name = "vid_window_x",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &vid_window_x },
};
static int vid_window_y;
static cvar_t vid_window_y_cvar = {
.name = "vid_window_y",
.description =
"",
.default_value = "0",
.flags = CVAR_ARCHIVE,
.value = { .type = &cexpr_int, .value = &vid_window_y },
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
//FIXME?int yeahimconsoled;
#define MAX_MODE_LIST 36
#define VID_ROW_SIZE 3
[win] Clean up a pile of messy externs
2021-03-31 01:36:30 +00:00
static DWORD WindowStyle, ExWindowStyle;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
[vid] Clean up windows video code a bit It turns out the dd and dib "driver" code is very specific to the software renderer. This does not fix the segfault on changing video mode, but I do know where the problem lies: the window is being destroyed and recreated without recreating the buffers. I suspect a clean solution to this will allow for window resizing in X as well.
2021-03-31 04:32:17 +00:00
int win_center_x, win_center_y;
RECT win_rect;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
DEVMODE win_gdevmode;
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
static qboolean startwindowed = 0;
//static qboolean windowed_mode_set;
//static int vid_fulldib_on_focus_mode;
static qboolean force_minimized;
static qboolean force_mode_set;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static qboolean vid_mode_set;
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
//static HICON hIcon;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
int vid_modenum = NO_MODE;
int vid_testingmode, vid_realmode;
double vid_testendtime;
int vid_default = MODE_WINDOWED;
static int windowed_default;
modestate_t modestate = MS_UNINIT;
byte vid_curpal[256 * 3];
int mode;
typedef struct {
modestate_t type;
int width;
int height;
int modenum;
int fullscreen;
char modedesc[13];
} vmode_t;
[vid] Clean up windows video code a bit It turns out the dd and dib "driver" code is very specific to the software renderer. This does not fix the segfault on changing video mode, but I do know where the problem lies: the window is being destroyed and recreated without recreating the buffers. I suspect a clean solution to this will allow for window resizing in X as well.
2021-03-31 04:32:17 +00:00
static vmode_t modelist[MAX_MODE_LIST] = {
{
.type = MS_WINDOWED,
.width = 320,
.height = 240,
.modedesc = "320x240",
.modenum = MODE_WINDOWED,
.fullscreen = 0,
},
{
.type = MS_WINDOWED,
.width = 640,
.height = 480,
.modedesc = "640x480",
.modenum = MODE_WINDOWED + 1,
.fullscreen = 0,
},
{
.type = MS_WINDOWED,
.width = 800,
.height = 600,
.modedesc = "800x600",
.modenum = MODE_WINDOWED + 2,
.fullscreen = 0,
}
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static int nummodes;
int aPage; // Current active display page
int vPage; // Current visible display page
int waitVRT = true; // True to wait for retrace on flip
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
static vmode_t badmode = {
.modedesc = "Bad mode",
};
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static int VID_SetMode (int modenum, const byte *palette);
static void __attribute__ ((used))
VID_RememberWindowPos (void)
{
RECT rect;
if (GetWindowRect (win_mainwindow, &rect)) {
if ((rect.left < GetSystemMetrics (SM_CXSCREEN)) &&
(rect.top < GetSystemMetrics (SM_CYSCREEN)) &&
(rect.right > 0) && (rect.bottom > 0)) {
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
vid_window_x = rect.left;
vid_window_y = rect.top;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
}
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#if 0
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static void
VID_CheckWindowXY (void)
{
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
if ((vid_window_x > (GetSystemMetrics (SM_CXSCREEN) - 160)) ||
(vid_window_y > (GetSystemMetrics (SM_CYSCREEN) - 120)) ||
(vid_window_x < 0) || (vid_window_y < 0)) {
vid_window_x = 0.0;
vid_window_y = 0.0;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#endif
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
void
Win_UpdateWindowStatus (int window_x, int window_y)
{
[vid] Clean up windows video code a bit It turns out the dd and dib "driver" code is very specific to the software renderer. This does not fix the segfault on changing video mode, but I do know where the problem lies: the window is being destroyed and recreated without recreating the buffers. I suspect a clean solution to this will allow for window resizing in X as well.
2021-03-31 04:32:17 +00:00
win_rect.left = window_x;
win_rect.top = window_y;
win_rect.right = window_x + viddef.width;
win_rect.bottom = window_y + viddef.height;
win_center_x = (win_rect.left + win_rect.right) / 2;
win_center_y = (win_rect.top + win_rect.bottom) / 2;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
IN_UpdateClipCursor ();
}
static qboolean
VID_CheckAdequateMem (int width, int height)
{
// there will always be enough ;)
return true;
}
static void
VID_InitModes (HINSTANCE hInstance)
{
WNDCLASS wc;
HDC hdc;
//FIXME hIcon = LoadIcon (hInstance, MAKEINTRESOURCE (IDI_ICON2));
/* Register the frame class */
wc.style = CS_OWNDC;
wc.lpfnWndProc = (WNDPROC) MainWndProc;
wc.cbClsExtra = 0;
wc.cbWndExtra = 0;
wc.hInstance = hInstance;
wc.hIcon = 0;
wc.hCursor = LoadCursor (NULL, IDC_ARROW);
wc.hbrBackground = NULL;
wc.lpszMenuName = 0;
wc.lpszClassName = "WinQuake";
if (!RegisterClass (&wc))
Sys_Error ("Couldn't register window class");
// automatically stretch the default mode up if > 640x480 desktop
// resolution
hdc = GetDC (NULL);
if ((GetDeviceCaps (hdc, HORZRES) > 800)
&& !COM_CheckParm ("-noautostretch")) {
vid_default = MODE_WINDOWED + 2;
} else if ((GetDeviceCaps (hdc, HORZRES) > 640)
&& !COM_CheckParm ("-noautostretch")) {
vid_default = MODE_WINDOWED + 1;
} else {
vid_default = MODE_WINDOWED;
}
// always start at the lowest mode then switch to the higher one if
// selected
vid_default = MODE_WINDOWED;
windowed_default = vid_default;
ReleaseDC (NULL, hdc);
nummodes = 3; // reserve space for windowed mode
}
static void
VID_GetDisplayModes (void)
{
DEVMODE devmode;
int i, modenum, existingmode, originalnummodes, lowestres;
BOOL stat;
// enumerate > 8 bpp modes
originalnummodes = nummodes;
modenum = 0;
lowestres = 99999;
do {
stat = EnumDisplaySettings (NULL, modenum, &devmode);
if ((devmode.dmPelsWidth <= MAXWIDTH)
&& (devmode.dmPelsHeight <= MAXHEIGHT)
&& (devmode.dmPelsWidth >= 320)
&& (devmode.dmPelsHeight >= 240)
&& (nummodes < MAX_MODE_LIST)) {
devmode.dmFields = DM_BITSPERPEL | DM_PELSWIDTH | DM_PELSHEIGHT;
if (ChangeDisplaySettings (&devmode, CDS_TEST | CDS_FULLSCREEN) ==
DISP_CHANGE_SUCCESSFUL) {
modelist[nummodes].type = MS_FULLDIB;
modelist[nummodes].width = devmode.dmPelsWidth;
modelist[nummodes].height = devmode.dmPelsHeight;
modelist[nummodes].modenum = 0;
modelist[nummodes].fullscreen = 1;
sprintf (modelist[nummodes].modedesc, "%dx%d",
(int) devmode.dmPelsWidth,
(int) devmode.dmPelsHeight);
// see is the mode already there
// (same dimensions but different refresh rate)
for (i = originalnummodes, existingmode = 0;
i < nummodes; i++) {
if ((modelist[nummodes].width == modelist[i].width)
&& (modelist[nummodes].height == modelist[i].height)) {
existingmode = 1;
break;
}
}
// if it's not add it to the list
if (!existingmode) {
if (modelist[nummodes].width < lowestres)
lowestres = modelist[nummodes].width;
nummodes++;
}
}
}
modenum++;
} while (stat);
if (nummodes != originalnummodes)
vid_default = MODE_FULLSCREEN_DEFAULT;
else
Sys_Printf ("No fullscreen DIB modes found\n");
}
void
Win_OpenDisplay (void)
{
VID_InitModes (global_hInstance);
VID_GetDisplayModes ();
vid_testingmode = 0;
// if (COM_CheckParm("-startwindowed"))
{
startwindowed = 1;
vid_default = windowed_default;
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#ifdef SPLASH_SCREEN
if (hwnd_dialog)
DestroyWindow (hwnd_dialog);
#endif
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
void
Win_CloseDisplay (void)
{
if (viddef.initialized) {
if (modestate == MS_FULLDIB) {
ChangeDisplaySettings (NULL, CDS_FULLSCREEN);
}
PostMessage (HWND_BROADCAST, WM_PALETTECHANGED, (WPARAM) win_mainwindow,
(LPARAM) 0);
PostMessage (HWND_BROADCAST, WM_SYSCOLORCHANGE, (WPARAM) 0, (LPARAM) 0);
[win] Clean up a pile of messy externs
2021-03-31 01:36:30 +00:00
Win_Activate (false, false);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
//FIXME? if (hwnd_dialog) DestroyWindow (hwnd_dialog);
if (win_mainwindow)
DestroyWindow (win_mainwindow);
vid_testingmode = 0;
viddef.initialized = false;
}
}
void
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
Win_SetVidMode (int width, int height)
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
{
//FIXME SCR_StretchInit();
force_mode_set = true;
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
//VID_SetMode (vid_default, palette);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
force_mode_set = false;
vid_realmode = vid_modenum;
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#if 0
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static void
VID_DestroyWindow (void)
{
if (modestate == MS_FULLDIB)
ChangeDisplaySettings (NULL, CDS_FULLSCREEN);
Win_UnloadAllDrivers ();
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#endif
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
static void
VID_CheckModedescFixup (int mode)
{
}
static qboolean
VID_SetWindowedMode (int modenum)
{
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#if 0
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
if (!windowed_mode_set) {
if (COM_CheckParm ("-resetwinpos")) {
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
vid_window_x = 0.0;
vid_window_y = 0.0;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
windowed_mode_set = true;
}
VID_CheckModedescFixup (modenum);
VID_DestroyWindow ();
DIBWidth = modelist[modenum].width;
DIBHeight = modelist[modenum].height;
WindowStyle = WS_OVERLAPPEDWINDOW | WS_CAPTION | WS_SYSMENU | WS_SIZEBOX |
WS_MINIMIZEBOX | WS_MAXIMIZEBOX | WS_CLIPSIBLINGS |
WS_CLIPCHILDREN | WS_THICKFRAME;
// WindowStyle = WS_OVERLAPPEDWINDOW|WS_VISIBLE;
ExWindowStyle = 0;
// the first time we're called to set the mode, create the window we'll use
// for the rest of the session
if (!vid_mode_set) {
} else {
SetWindowLong (win_mainwindow, GWL_STYLE, WindowStyle | WS_VISIBLE);
SetWindowLong (win_mainwindow, GWL_EXSTYLE, ExWindowStyle);
}
if (!SetWindowPos (win_mainwindow,
NULL,
0, 0,
WindowRect.right - WindowRect.left,
WindowRect.bottom - WindowRect.top,
SWP_NOCOPYBITS | SWP_NOZORDER | SWP_HIDEWINDOW)) {
Sys_Error ("Couldn't resize DIB window");
}
// position and show the DIB window
VID_CheckWindowXY ();
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
SetWindowPos (win_mainwindow, NULL, vid_window_x,
vid_window_y, 0, 0,
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
SWP_NOSIZE | SWP_NOZORDER | SWP_SHOWWINDOW | SWP_DRAWFRAME);
if (force_minimized)
ShowWindow (win_mainwindow, SW_MINIMIZE);
else
ShowWindow (win_mainwindow, SW_SHOWDEFAULT);
UpdateWindow (win_mainwindow);
modestate = MS_WINDOWED;
vid_fulldib_on_focus_mode = 0;
viddef.numpages = 1;
// viddef.height = viddef.conheight = DIBHeight;
// viddef.width = viddef.conwidth = DIBWidth;
viddef.height = viddef.conheight = DIBHeight;
viddef.width = viddef.conwidth = DIBWidth;
//FIXME? if (!yeahimconsoled){
//FIXME? viddef.vconheight = DIBHeight;
//FIXME? viddef.vconwidth = DIBWidth;
//FIXME? }
SendMessage (win_mainwindow, WM_SETICON, (WPARAM) TRUE, (LPARAM) hIcon);
SendMessage (win_mainwindow, WM_SETICON, (WPARAM) FALSE, (LPARAM) hIcon);
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#endif
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
return true;
}
static qboolean
VID_SetFullDIBMode (int modenum)
{
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#if 0
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
VID_DestroyWindow ();
win_gdevmode.dmFields = DM_PELSWIDTH | DM_PELSHEIGHT;
win_gdevmode.dmPelsWidth = modelist[modenum].width;
win_gdevmode.dmPelsHeight = modelist[modenum].height;
win_gdevmode.dmSize = sizeof (win_gdevmode);
if (ChangeDisplaySettings (&win_gdevmode, CDS_FULLSCREEN) !=
DISP_CHANGE_SUCCESSFUL)
Sys_Error ("Couldn't set fullscreen DIB mode");
modestate = MS_FULLDIB;
vid_fulldib_on_focus_mode = modenum;
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
WindowRect.top = WindowRect.left = 0;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
WindowRect.right = modelist[modenum].width;
WindowRect.bottom = modelist[modenum].height;
DIBWidth = modelist[modenum].width;
DIBHeight = modelist[modenum].height;
WindowStyle = WS_POPUP | WS_SYSMENU | WS_CLIPSIBLINGS | WS_CLIPCHILDREN;
ExWindowStyle = 0;
AdjustWindowRectEx (&WindowRect, WindowStyle, FALSE, 0);
SetWindowLong (win_mainwindow, GWL_STYLE, WindowStyle | WS_VISIBLE);
SetWindowLong (win_mainwindow, GWL_EXSTYLE, ExWindowStyle);
if (!SetWindowPos (win_mainwindow,
NULL,
0, 0,
WindowRect.right - WindowRect.left,
WindowRect.bottom - WindowRect.top,
SWP_NOCOPYBITS | SWP_NOZORDER)) {
Sys_Error ("Couldn't resize DIB window");
}
// position and show the DIB window
SetWindowPos (win_mainwindow, HWND_TOPMOST, 0, 0, 0, 0,
SWP_NOSIZE | SWP_SHOWWINDOW | SWP_DRAWFRAME);
ShowWindow (win_mainwindow, SW_SHOWDEFAULT);
UpdateWindow (win_mainwindow);
viddef.numpages = 1;
#ifdef SCALED2D
viddef.height = viddef.conheight = DIBHeight;
viddef.width = viddef.conwidth = DIBWidth;
// viddef.vconwidth = 320;
// viddef.vconheight = 200;
//FIXME? if (!yeahimconsoled){
//FIXME? viddef.vconheight = DIBHeight;
//FIXME? viddef.vconwidth = DIBWidth;
//FIXME? }
#else
viddef.height = viddef.conheight = DIBHeight;
viddef.width = viddef.conwidth = DIBWidth;
#endif
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
#endif
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
return true;
}
static void
VID_RestoreOldMode (int original_mode)
{
static qboolean inerror = false;
if (inerror)
return;
inerror = true;
// make sure mode set happens (video mode changes)
vid_modenum = original_mode - 1;
if (!VID_SetMode (original_mode, vid_curpal)) {
vid_modenum = MODE_WINDOWED - 1;
if (!VID_SetMode (windowed_default, vid_curpal))
Sys_Error ("Can't set any video mode");
}
inerror = false;
}
static void __attribute__ ((used))
VID_SetDefaultMode (void)
{
if (viddef.initialized)
VID_SetMode (0, vid_curpal);
IN_DeactivateMouse ();
}
static vmode_t *
VID_GetModePtr (int modenum)
{
if ((modenum >= 0) && (modenum < nummodes))
return &modelist[modenum];
else
return &badmode;
}
static char *
VID_GetModeDescription (int mode)
{
char *pinfo;
vmode_t *pv;
if ((mode < 0) || (mode >= nummodes))
return NULL;
VID_CheckModedescFixup (mode);
pv = VID_GetModePtr (mode);
pinfo = pv->modedesc;
return pinfo;
}
static int
VID_SetMode (int modenum, const byte *palette)
{
int original_mode; // FIXME, temp;
qboolean stat;
MSG msg;
HDC hdc;
while ((modenum >= nummodes) || (modenum < 0)) {
if (vid_modenum == NO_MODE) {
if (modenum == vid_default) {
modenum = windowed_default;
} else {
modenum = vid_default;
}
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
vid_mode = modenum;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
} else {
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
vid_mode = vid_modenum;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
return 0;
}
}
if (!force_mode_set && (modenum == vid_modenum))
return true;
//FIXME CDAudio_Pause ();
//FIXME S_ClearBuffer ();
if (vid_modenum == NO_MODE)
original_mode = windowed_default;
else
original_mode = vid_modenum;
// Set either the fullscreen or windowed mode
if (modelist[modenum].type == MS_WINDOWED) {
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
if (_windowed_mouse) {
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
stat = VID_SetWindowedMode (modenum);
IN_ActivateMouse ();
IN_HideMouse ();
} else {
IN_DeactivateMouse ();
IN_ShowMouse ();
stat = VID_SetWindowedMode (modenum);
}
} else {
stat = VID_SetFullDIBMode (modenum);
IN_ActivateMouse ();
IN_HideMouse ();
}
Win_UpdateWindowStatus (0, 0); // FIXME right numbers?
//FIXME CDAudio_Resume ();
if (!stat) {
VID_RestoreOldMode (original_mode);
return false;
}
// now we try to make sure we get the focus on the mode switch, because
// sometimes in some systems we don't. We grab the foreground, then
// finish setting up, pump all our messages, and sleep for a little while
// to let messages finish bouncing around the system, then we put
// ourselves at the top of the z order, then grab the foreground again,
// Who knows if it helps, but it probably doesn't hurt
if (!force_minimized)
SetForegroundWindow (win_mainwindow);
hdc = GetDC (NULL);
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
if (GetDeviceCaps (hdc, RASTERCAPS) & RC_PALETTE) {
[win] Clean up a pile of messy externs
2021-03-31 01:36:30 +00:00
win_palettized = true;
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
} else {
[win] Clean up a pile of messy externs
2021-03-31 01:36:30 +00:00
win_palettized = false;
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
}
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
ReleaseDC (NULL, hdc);
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
vid_modenum = modenum;
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
vid_mode = vid_modenum;
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
while (PeekMessage (&msg, NULL, 0, 0, PM_REMOVE)) {
TranslateMessage (&msg);
DispatchMessage (&msg);
}
Sleep (100);
if (!force_minimized) {
SetWindowPos (win_mainwindow, HWND_TOP, 0, 0, 0, 0,
SWP_DRAWFRAME | SWP_NOMOVE | SWP_NOSIZE | SWP_SHOWWINDOW |
SWP_NOCOPYBITS);
SetForegroundWindow (win_mainwindow);
}
// fix the leftover Alt from any Alt-Tab or the like that switched us away
[win] Update windows input for new binding system Seems to work nicely for keyboard (though key bindings are not cross-platform). Mouse not tested yet, and I expect there are problems with it for absolute inputs (yay mouse warp :P).
2021-11-19 04:52:40 +00:00
IN_ClearStates ();
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
Sys_Printf ("%s\n", VID_GetModeDescription (vid_modenum));
viddef.recalc_refdef = 1;
//FIXME SCR_StretchInit();
//FIXME SCR_StretchRefresh();
//FIXME SCR_CvarCheck();
return true;
}
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
void
Win_CreateWindow (int width, int height)
{
RECT rect = {
.top = 0,
.left = 0,
.right = width,
.bottom = height,
};
AdjustWindowRectEx (&rect, WindowStyle, FALSE, 0);
// sound initialization has to go here, preceded by a windowed mode set,
// so there's a window for DirectSound to work with but we're not yet
// fullscreen so the "hardware already in use" dialog is visible if it
// gets displayed
// keep the window minimized until we're ready for the first real mode set
win_mainwindow = CreateWindowEx (ExWindowStyle,
"WinQuake",
"WinQuake",
WindowStyle,
0, 0,
rect.right - rect.left,
rect.bottom - rect.top,
NULL, NULL, global_hInstance, NULL);
if (!win_mainwindow)
Sys_Error ("Couldn't create DIB window");
// done
vid_mode_set = true;
//FIXME if (firsttime) S_Init ();
Win_UpdateWindowStatus (0, 0); // FIXME right numbers?
HDC hdc = GetDC (NULL);
if (GetDeviceCaps (hdc, RASTERCAPS) & RC_PALETTE) {
win_palettized = true;
} else {
win_palettized = false;
}
ReleaseDC (NULL, hdc);
//vid_modenum = modenum;
//Cvar_SetValue (vid_mode, (float) vid_modenum);
MSG msg;
while (PeekMessage (&msg, NULL, 0, 0, PM_REMOVE)) {
TranslateMessage (&msg);
DispatchMessage (&msg);
}
Sleep (100);
if (!force_minimized) {
SetWindowPos (win_mainwindow, HWND_TOP, 0, 0, 0, 0,
SWP_DRAWFRAME | SWP_NOMOVE | SWP_NOSIZE | SWP_SHOWWINDOW |
SWP_NOCOPYBITS);
SetForegroundWindow (win_mainwindow);
}
// fix the leftover Alt from any Alt-Tab or the like that switched us away
[win] Update windows input for new binding system Seems to work nicely for keyboard (though key bindings are not cross-platform). Mouse not tested yet, and I expect there are problems with it for absolute inputs (yay mouse warp :P).
2021-11-19 04:52:40 +00:00
IN_ClearStates ();
[vid] Make windows vid init conform to QF's method Ie, respecting cvars.
2021-03-31 06:50:10 +00:00
Sys_Printf ("%s\n", VID_GetModeDescription (vid_modenum));
viddef.recalc_refdef = 1;
}
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
//==========================================================================
/*
================
VID_HandlePause
================
*/
static void __attribute__ ((used))
VID_HandlePause (qboolean pause)
{
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
if ((modestate == MS_WINDOWED) && _windowed_mouse) {
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
if (pause) {
IN_DeactivateMouse ();
IN_ShowMouse ();
} else {
IN_ActivateMouse ();
IN_HideMouse ();
}
}
}
/*
===================================================================
MAIN WINDOW
===================================================================
*/
typedef struct {
int modenum;
char *desc;
int iscur;
int width;
} modedesc_t;
#define MAX_COLUMN_SIZE 5
#define MODE_AREA_HEIGHT (MAX_COLUMN_SIZE + 6)
#define MAX_MODEDESCS (MAX_COLUMN_SIZE*3)
//static modedesc_t modedescs[MAX_MODEDESCS];
static int
VID_NumModes (void)
{
return nummodes;
}
static char * __attribute__((used))
VID_GetModeDescriptionMemCheck (int mode)
{
char *pinfo;
vmode_t *pv;
if ((mode < 0) || (mode >= nummodes))
return NULL;
VID_CheckModedescFixup (mode);
pv = VID_GetModePtr (mode);
pinfo = pv->modedesc;
if (VID_CheckAdequateMem (pv->width, pv->height)) {
return pinfo;
} else {
return NULL;
}
}
// Tacks on "windowed" or "fullscreen"
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
static const char * __attribute__((used))
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
VID_GetModeDescription2 (int mode)
{
vmode_t *pv;
if ((mode < 0) || (mode >= nummodes))
return NULL;
VID_CheckModedescFixup (mode);
pv = VID_GetModePtr (mode);
if (modelist[mode].type == MS_FULLSCREEN) {
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
return va (0, "%s fullscreen", pv->modedesc);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
} else if (modelist[mode].type == MS_FULLDIB) {
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
return va (0, "%s fullscreen", pv->modedesc);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
} else {
[win] Fix a handful of small bugs for windows Only 64-bit windows is tested, and there are still various failures, but QF is limping along in windows again. nq-sdl works for sw, and sw32, gl and glsl are mostly black (but not entirely for gl?), vulkan is not supported with sdl. nq-win works for sw and sw32, and sort of for vulkan (very dark and upside-down?). gl and glsl complain about vid mode, qw-client-[sdl,win] seem to be the same, but something is wrong with the console (reading keyboard input).
2021-03-30 11:09:13 +00:00
return va (0, "%s windowed", pv->modedesc);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
}
// KJB: Added this to return the mode driver name in description for console
static char *
VID_GetExtModeDescription (int mode)
{
static char pinfo[40];
vmode_t *pv;
if ((mode < 0) || (mode >= nummodes))
return NULL;
VID_CheckModedescFixup (mode);
pv = VID_GetModePtr (mode);
if (modelist[mode].type == MS_FULLDIB) {
sprintf (pinfo, "%s fullscreen", pv->modedesc);
} else {
sprintf (pinfo, "%s windowed", pv->modedesc);
}
return pinfo;
}
static void
VID_DescribeCurrentMode_f (void)
{
Sys_Printf ("%s\n", VID_GetExtModeDescription (vid_modenum));
}
static void
VID_NumModes_f (void)
{
if (nummodes == 1)
Sys_Printf ("%d video mode is available\n", nummodes);
else
Sys_Printf ("%d video modes are available\n", nummodes);
}
static void
VID_DescribeMode_f (void)
{
int modenum;
modenum = atoi (Cmd_Argv (1));
Sys_Printf ("%s\n", VID_GetExtModeDescription (modenum));
}
static void
VID_DescribeModes_f (void)
{
int i, lnummodes;
char *pinfo;
qboolean na;
vmode_t *pv;
na = false;
lnummodes = VID_NumModes ();
for (i = 0; i < lnummodes; i++) {
pv = VID_GetModePtr (i);
pinfo = VID_GetExtModeDescription (i);
if (VID_CheckAdequateMem (pv->width, pv->height)) {
Sys_Printf ("%2d: %s\n", i, pinfo);
} else {
Sys_Printf ("**: %s\n", pinfo);
na = true;
}
}
if (na) {
Sys_Printf ("\n[**: not enough system RAM for mode]\n");
}
}
static void
VID_TestMode_f (void)
{
int modenum;
double testduration;
if (!vid_testingmode) {
modenum = atoi (Cmd_Argv (1));
if (VID_SetMode (modenum, vid_curpal)) {
vid_testingmode = 1;
testduration = atof (Cmd_Argv (2));
if (testduration == 0)
testduration = 5.0;
vid_testendtime = Sys_DoubleTime () + testduration;
}
}
}
static void
VID_Windowed_f (void)
{
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
VID_SetMode (vid_windowed_mode, vid_curpal);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
static void
VID_Fullscreen_f (void)
{
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
VID_SetMode (vid_fullscreen_mode, vid_curpal);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
}
static void
VID_Minimize_f (void)
{
// we only support minimizing windows; if you're fullscreen,
// switch to windowed first
if (modestate == MS_WINDOWED)
ShowWindow (win_mainwindow, SW_MINIMIZE);
}
static void
VID_ForceMode_f (void)
{
int modenum;
if (!vid_testingmode) {
modenum = atoi (Cmd_Argv (1));
force_mode_set = 1;
VID_SetMode (modenum, vid_curpal);
force_mode_set = 0;
}
}
void
Win_SetCaption (const char *text)
{
if (win_mainwindow) {
SetWindowText (win_mainwindow, text);
}
}
//static WORD systemgammaramps[3][256];
static WORD currentgammaramps[3][256];
qboolean
Win_SetGamma (double gamma)
{
int i;
HDC hdc = GetDC (NULL);
for (i = 0; i < 256; i++) {
currentgammaramps[2][i] = currentgammaramps[1][i] =
currentgammaramps[0][i] = viddef.gammatable[i] * 256;
}
i = SetDeviceGammaRamp (hdc, &currentgammaramps[0][0]);
ReleaseDC (NULL, hdc);
return i;
}
void
Win_Init_Cvars (void)
{
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
Cvar_Register (&vid_ddraw_cvar, 0, 0);
Cvar_Register (&vid_mode_cvar, 0, 0);
Cvar_Register (&vid_wait_cvar, 0, 0);
Cvar_Register (&vid_nopageflip_cvar, 0, 0);
Cvar_Register (&_vid_wait_override_cvar, 0, 0);
Cvar_Register (&_vid_default_mode_cvar, 0, 0);
Cvar_Register (&_vid_default_mode_win_cvar, 0, 0);
Cvar_Register (&vid_config_x_cvar, 0, 0);
Cvar_Register (&vid_config_y_cvar, 0, 0);
Cvar_Register (&vid_stretch_by_2_cvar, 0, 0);
Cvar_Register (&_windowed_mouse_cvar, 0, 0);
Cvar_Register (&vid_fullscreen_mode_cvar, 0, 0);
Cvar_Register (&vid_windowed_mode_cvar, 0, 0);
Cvar_Register (&block_switch_cvar, 0, 0);
Cvar_Register (&vid_window_x_cvar, 0, 0);
Cvar_Register (&vid_window_y_cvar, 0, 0);
Get QF cross-compiling using MXE/mingw32 This includes -win clients (no clue if anything actually works yet).
2021-03-27 11:09:37 +00:00
Cmd_AddCommand ("vid_testmode", VID_TestMode_f, "");
Cmd_AddCommand ("vid_nummodes", VID_NumModes_f, "");
Cmd_AddCommand ("vid_describecurrentmode", VID_DescribeCurrentMode_f, "");
Cmd_AddCommand ("vid_describemode", VID_DescribeMode_f, "");
Cmd_AddCommand ("vid_describemodes", VID_DescribeModes_f, "");
Cmd_AddCommand ("vid_forcemode", VID_ForceMode_f, "");
Cmd_AddCommand ("vid_windowed", VID_Windowed_f, "");
Cmd_AddCommand ("vid_fullscreen", VID_Fullscreen_f, "");
Cmd_AddCommand ("vid_minimize", VID_Minimize_f, "");
}
[win] Hook up windows input Not sure it actually works yet (can't test right now), but since the x11 switch-over worked, it should.
2021-09-28 03:48:47 +00:00
extern int win_force_link;
static __attribute__((used)) int *context_win_force_link = &win_force_link;
Reference in a new issue Copy permalink