taisei/src/menu/options.c

1981 lines
48 KiB
C
Raw Permalink Normal View History

/*
* This software is licensed under the terms of the MIT License.
* See COPYING for further information.
* ---
2024-05-16 23:30:41 +02:00
* Copyright (c) 2011-2024, Lukas Weber <laochailan@web.de>.
* Copyright (c) 2012-2024, Andrei Alexeyev <akari@taisei-project.org>.
*/
#include "options.h"
2017-02-12 04:43:52 +01:00
#include "common.h"
#include "config.h"
#include "events.h"
2019-04-20 18:06:27 +02:00
#include "mainmenu.h"
#include "menu.h"
#include "audio/audio.h"
#include "gamepad.h"
#include "renderer/api.h"
#include "resource/font.h"
#include "util/graphics.h"
#include "video.h"
#define OPTIONS_ACTIVE_X_OFFSET 20 /* FIXME hardcoded in draw_menu_list */
#define OPTIONS_ITEM_HEIGHT 20 /* FIXME hardcoded in draw_menu_list */
#define OPTIONS_X_MARGIN 100
#define OPTIONS_Y_MARGIN 100
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
typedef struct OptionBinding OptionBinding;
typedef int (*BindingGetter)(OptionBinding*);
typedef int (*BindingSetter)(OptionBinding*, int);
typedef bool (*BindingDependence)(void);
typedef enum BindingType {
BT_IntValue,
BT_KeyBinding,
BT_StrValue,
BT_Resolution,
BT_FramebufferResolution,
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
BT_Scale,
BT_GamepadKeyBinding,
BT_GamepadAxisBinding,
BT_GamepadDevice,
BT_VideoDisplay,
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
} BindingType;
typedef enum OptionsMenuEntryArgType {
ARGTYPE_BIND = 0xafafaf00,
ARGTYPE_OTHER,
} OptionsMenuEntryArgType;
// Shared header for all structs used as args in menu entries,
// because past me had the brilliant idea of storing the bindings in there
// and now there is a need to tell them apart from other stuff.
// One of these days I'm gonna rewrite this whole file from scratch (copium)
// along with the entire menu system (mega copium), but that day is not today.
typedef struct OptionsMenuArgHeader {
OptionsMenuEntryArgType type;
} OptionsMenuArgHeader;
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
typedef struct OptionBinding {
OptionsMenuArgHeader header;
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
union {
char **values;
char *strvalue;
};
bool displaysingle;
int valrange_min;
int valrange_max;
float scale_min;
float scale_max;
float scale_step;
BindingGetter getter;
BindingSetter setter;
BindingDependence dependence;
int selected;
int configentry;
BindingType type;
bool blockinput;
int pad;
} OptionBinding;
// --- Menu entry <-> config option binding stuff --- //
static OptionBinding* bind_new(void) {
return ALLOC(OptionBinding, {
.header.type = ARGTYPE_BIND,
.selected = -1,
.configentry = -1,
});
}
static void bind_free(OptionBinding *bind) {
2012-08-16 18:15:46 +02:00
int i;
if(bind->type == BT_StrValue) {
mem_free(bind->strvalue);
} else if(bind->values) {
assert(bind->valrange_min == 0);
for(i = 0; i <= bind->valrange_max; ++i) {
mem_free(*(bind->values+i));
}
mem_free(bind->values);
}
2012-08-16 18:15:46 +02:00
}
static OptionBinding* bind_get(MenuData *m, int idx) {
MenuEntry *e = dynarray_get_ptr(&m->entries, idx);
OptionsMenuArgHeader *header = e->arg;
if(!header) {
return NULL;
}
if(header->type == ARGTYPE_BIND) {
return CASTPTR_ASSUME_ALIGNED(header, OptionBinding);
}
assume(header->type == ARGTYPE_OTHER);
return NULL;
}
// BT_IntValue: integer and boolean options
// Values are defined with bind_addvalue or bind_setrange
static OptionBinding* bind_option(int cfgentry, BindingGetter getter, BindingSetter setter) {
2012-08-16 23:35:48 +02:00
OptionBinding *bind = bind_new();
bind->type = BT_IntValue;
bind->configentry = cfgentry;
bind->getter = getter;
bind->setter = setter;
return bind;
}
// BT_KeyBinding: keyboard action mapping options
static OptionBinding* bind_keybinding(int cfgentry) {
2012-08-16 23:35:48 +02:00
OptionBinding *bind = bind_new();
bind->configentry = cfgentry;
bind->type = BT_KeyBinding;
return bind;
}
// BT_GamepadKeyBinding: gamepad action mapping options
static OptionBinding* bind_gpbinding(int cfgentry) {
OptionBinding *bind = bind_new();
bind->configentry = cfgentry;
bind->type = BT_GamepadKeyBinding;
return bind;
}
2017-02-27 23:58:47 +01:00
// BT_GamepadAxisBinding: gamepad axis mapping options
attr_unused
static OptionBinding* bind_gpaxisbinding(int cfgentry) {
2017-02-27 23:58:47 +01:00
OptionBinding *bind = bind_new();
bind->configentry = cfgentry;
bind->type = BT_GamepadAxisBinding;
return bind;
}
static int bind_gpdev_get(OptionBinding *b) {
const char *guid = config_get_str(b->configentry);
int val = gamepad_device_num_from_guid(guid);
if(val == GAMEPAD_DEVNUM_ANY) {
val = -1;
}
return val;
}
static int bind_gpdev_set(OptionBinding *b, int v) {
char guid[33] = {0};
if(v == -1) {
v = GAMEPAD_DEVNUM_ANY;
}
gamepad_device_guid(v, guid, sizeof(guid));
if(*guid) {
config_set_str(b->configentry, guid);
if(v == GAMEPAD_DEVNUM_ANY) {
v = -1;
}
b->selected = v;
}
return b->selected;
}
#ifndef __SWITCH__
// BT_GamepadDevice: dynamic device list
static OptionBinding* bind_gpdevice(int cfgentry) {
OptionBinding *bind = bind_new();
bind->configentry = cfgentry;
bind->type = BT_GamepadDevice;
bind->getter = bind_gpdev_get;
bind->setter = bind_gpdev_set;
bind->valrange_min = -1;
bind->valrange_max = 0; // updated later
bind->selected = gamepad_device_num_from_guid(config_get_str(bind->configentry));
return bind;
}
// BT_StrValue: with a half-assed "textbox"
static OptionBinding* bind_stroption(ConfigIndex cfgentry) {
2012-08-16 23:35:48 +02:00
OptionBinding *bind = bind_new();
bind->type = BT_StrValue;
2012-08-16 23:35:48 +02:00
bind->configentry = cfgentry;
stralloc(&bind->strvalue, config_get_str(cfgentry));
return bind;
}
#endif
// BT_Resolution: super-special binding type for the resolution setting
static void bind_resolution_update(OptionBinding *bind) {
// FIXME This is brittle. The least we could do is to explicitly store whether the currently selected value represents a fullscreen index or windowed.
bool fullscreen = video_is_fullscreen();
uint nmodes = video_get_num_modes(fullscreen);
VideoMode cur = video_get_current_mode();
log_debug("Fullscreen: %i", fullscreen);
log_debug("Prev selected: %i", bind->selected);
bind->valrange_min = 0;
bind->valrange_max = nmodes - 1;
bind->selected = -1;
log_debug("nmodes = %i", nmodes);
for(int i = 0; i < nmodes; ++i) {
2023-06-20 00:33:01 +02:00
VideoMode m;
attr_unused bool ok = video_get_mode(i, fullscreen, &m);
assert(ok);
log_debug("#%i %ix%i", i, m.width, m.height);
if(m.width == cur.width && m.height == cur.height) {
2017-03-01 01:40:50 +01:00
bind->selected = i;
log_debug("selected #%i", bind->selected);
2017-03-01 01:40:50 +01:00
}
}
}
2017-03-01 01:40:50 +01:00
static OptionBinding* bind_resolution(void) {
OptionBinding *bind = bind_new();
bind->type = BT_Resolution;
bind->selected = -1;
bind_resolution_update(bind);
2012-07-28 22:53:53 +02:00
return bind;
}
// BT_FramebufferResolution: not an actual setting (yet); just display effective resolution in pixels
// This may be different from BT_Resolution in the high-DPI case
// (BT_Resolution is a misnomer; it represents the window size in screen-space units)
static OptionBinding* bind_fb_resolution(void) {
OptionBinding *bind = bind_new();
bind->type = BT_FramebufferResolution;
return bind;
}
// BT_Scale: float values clamped to a range
static OptionBinding* bind_scale(int cfgentry, float smin, float smax, float step) {
2012-08-16 23:35:48 +02:00
OptionBinding *bind = bind_new();
bind->type = BT_Scale;
bind->configentry = cfgentry;
2012-08-16 23:35:48 +02:00
bind->scale_min = smin;
bind->scale_max = smax;
bind->scale_step = step;
2012-08-16 23:35:48 +02:00
return bind;
}
2011-07-03 21:52:33 +02:00
// Returns a pointer to the first found binding that blocks input. If none found, returns NULL.
static OptionBinding* bind_getinputblocking(MenuData *m) {
dynarray_foreach_idx(&m->entries, int i, {
2012-08-16 18:15:46 +02:00
OptionBinding *bind = bind_get(m, i);
if(bind && bind->blockinput) {
2012-08-16 18:15:46 +02:00
return bind;
}
});
return NULL;
}
2011-07-03 21:52:33 +02:00
// Adds a value to a BT_IntValue type binding
static int bind_addvalue(OptionBinding *b, char *val) {
assert(b->valrange_min == 0);
if(b->values == NULL) {
b->values = mem_alloc(sizeof(char*));
b->valrange_min = 0;
b->valrange_max = 0;
} else {
assert(b->valrange_min == 0);
++b->valrange_max;
}
b->values = mem_realloc(b->values, (1 + b->valrange_max) * sizeof(char*));
b->values[b->valrange_max] = mem_strdup(val);
return b->valrange_max;
}
attr_unused
static void bind_setvaluerange(OptionBinding *b, int vmin, int vmax) {
assert(b->values == NULL);
b->valrange_min = vmin;
b->valrange_max = vmax;
}
2011-07-03 21:52:33 +02:00
// Called to select a value of a BT_IntValue type binding by index
static int bind_setvalue(OptionBinding *b, int v) {
2012-07-28 22:53:53 +02:00
if(b->setter)
return b->selected = b->setter(b, v);
else
return b->selected = v;
}
2011-07-03 21:52:33 +02:00
// Called to get the selected value of a BT_IntValue type binding by index
static int bind_getvalue(OptionBinding *b) {
2012-08-16 23:35:48 +02:00
if(b->getter) {
b->selected = b->getter(b);
2012-08-16 23:35:48 +02:00
}
2012-08-16 23:35:48 +02:00
return b->selected;
}
2011-07-03 21:52:33 +02:00
// Selects the next to current value of a BT_IntValue type binding
static int bind_setnext(OptionBinding *b) {
int s = b->selected + 1;
if(s > b->valrange_max) {
s = b->valrange_min;
}
return bind_setvalue(b, s);
}
2011-07-03 21:52:33 +02:00
// Selects the previous to current value of a BT_IntValue type binding
static int bind_setprev(OptionBinding *b) {
int s = b->selected - 1;
if(s < b->valrange_min) {
s = b->valrange_max;
}
return bind_setvalue(b, s);
}
static bool bind_isactive(OptionBinding *b) {
if(!b->dependence)
return true;
return b->dependence();
}
// --- Shared binding callbacks --- //
static int bind_common_onoff_get(OptionBinding *b) {
return !config_get_int(b->configentry);
}
static int bind_common_onoff_set(OptionBinding *b, int v) {
return !config_set_int(b->configentry, !v);
}
static int bind_common_onoff_inverted_get(OptionBinding *b) {
return config_get_int(b->configentry);
}
static int bind_common_onoff_inverted_set(OptionBinding *b, int v) {
return config_set_int(b->configentry, v);
}
static int bind_common_onoffplus_get(OptionBinding *b) {
int v = config_get_int(b->configentry);
if(v > 1)
return v;
return !v;
}
static int bind_common_onoffplus_set(OptionBinding *b, int v) {
if(v > 1)
return config_set_int(b->configentry, v);
return !config_set_int(b->configentry, !v);
}
#define bind_common_int_get bind_common_onoff_inverted_get
#define bind_common_int_set bind_common_onoff_inverted_set
// BT_VideoDisplay: fullscreen display number
static OptionBinding* bind_video_display(int cfgentry) {
OptionBinding *bind = bind_new();
bind->configentry = cfgentry;
bind->type = BT_VideoDisplay;
bind->getter = bind_common_int_get;
bind->setter = bind_common_int_set;
bind->valrange_min = 0;
bind->valrange_max = 0; // updated later
bind->selected = video_current_display();
return bind;
}
static int bind_common_intplus1_get(OptionBinding *b) {
2017-12-26 12:07:40 +01:00
return config_get_int(b->configentry) - 1;
}
static int bind_common_intplus1_set(OptionBinding *b, int v) {
2017-12-26 12:07:40 +01:00
return config_set_int(b->configentry, v + 1) - 1;
}
// --- Binding callbacks for individual options --- //
static bool bind_resizable_dependence(void) {
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
return video_query_capability(VIDEO_CAP_EXTERNAL_RESIZE) == VIDEO_AVAILABLE;
}
static bool bind_bgquality_dependence(void) {
return !config_get_int(CONFIG_NO_STAGEBG);
}
static bool bind_resolution_dependence(void) {
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
return video_query_capability(VIDEO_CAP_CHANGE_RESOLUTION) == VIDEO_AVAILABLE;
}
static bool bind_fb_resolution_dependence(void) {
return false;
}
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
static bool bind_fullscreen_dependence(void) {
return video_query_capability(VIDEO_CAP_FULLSCREEN) == VIDEO_AVAILABLE;
}
static int bind_resolution_set(OptionBinding *b, int v) {
2023-06-20 00:33:01 +02:00
VideoMode m;
if(video_get_mode(v, video_is_fullscreen(), &m)) {
config_set_int(CONFIG_VID_WIDTH, m.width);
config_set_int(CONFIG_VID_HEIGHT, m.height);
2017-03-01 01:40:50 +01:00
}
return v;
}
static int bind_power_set(OptionBinding *b, int v) {
return config_set_int(b->configentry, v * 100) / 100;
}
static int bind_power_get(OptionBinding *b) {
return config_get_int(b->configentry) / 100;
}
static int bind_gamepad_set(OptionBinding *b, int v) {
v = bind_common_onoff_set(b, v);
if(v == 0) {
gamepad_init();
}
return v;
}
// --- Creating, destroying, filling the menu --- //
typedef struct ConfirmDialog {
OptionsMenuArgHeader header;
const char *const text;
void (*const action)(void);
} ConfirmDialog;
static const ConfirmDialog dialog_reset_default = {
.header.type = ARGTYPE_OTHER,
.text = "Reset all settings to defaults?",
.action = config_reset,
};
static const ConfirmDialog dialog_reset_saved = {
.header.type = ARGTYPE_OTHER,
.text = "Reset all settings to the last saved values?",
.action = config_load,
};
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
typedef struct OptionsMenuContext {
const char *title;
void *data;
MenuData *submenu;
MenuData *submenu_fading;
ConfirmDialog *confirm_dialog;
float submenu_alpha;
void (*draw_overlay)(MenuData *m, struct OptionsMenuContext *ctx);
struct {
bool allowed;
bool active;
} gamepad_testmode;
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
} OptionsMenuContext;
static void destroy_options_menu(MenuData *m) {
bool change_vidmode = false;
dynarray_foreach_idx(&m->entries, int i, {
2012-08-16 18:15:46 +02:00
OptionBinding *bind = bind_get(m, i);
2017-12-28 02:35:53 +01:00
if(!bind) {
2012-08-16 18:15:46 +02:00
continue;
2017-12-28 02:35:53 +01:00
}
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
if(bind->type == BT_Resolution && video_query_capability(VIDEO_CAP_CHANGE_RESOLUTION) == VIDEO_AVAILABLE) {
2012-08-16 18:15:46 +02:00
if(bind->selected != -1) {
bool fullscreen = video_is_fullscreen();
2023-06-20 00:33:01 +02:00
VideoMode mode;
if(video_get_mode(bind->selected, fullscreen, &mode)) {
config_set_int(CONFIG_VID_WIDTH, mode.width);
config_set_int(CONFIG_VID_HEIGHT, mode.height);
change_vidmode = true;
}
2012-07-28 22:53:53 +02:00
}
}
2012-08-16 18:15:46 +02:00
bind_free(bind);
mem_free(bind);
});
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
if(change_vidmode) {
video_set_mode(
config_get_int(CONFIG_VID_DISPLAY),
config_get_int(CONFIG_VID_WIDTH),
config_get_int(CONFIG_VID_HEIGHT),
config_get_int(CONFIG_FULLSCREEN),
config_get_int(CONFIG_VID_RESIZABLE)
);
}
Emscripten compatibility (#161) * Major refactoring of the main loop(s) and control flow (WIP) run_at_fps() is gone 🦀 Instead of nested blocking event loops, there is now an eventloop API that manages an explicit stack of scenes. This makes Taisei a lot more portable to async environments where spinning a loop forever without yielding control simply is not an option, and that is the entire point of this change. A prime example of such an environment is the Web (via emscripten). Taisei was able to run there through a terrible hack: inserting emscripten_sleep calls into the loop, which would yield to the browser. This has several major drawbacks: first of all, every function that could possibly call emscripten_sleep must be compiled into a special kind of bytecode, which then has to be interpreted at runtime, *much* slower than JITed WebAssembly. And that includes *everything* down the call stack, too! For more information, see https://emscripten.org/docs/porting/emterpreter.html Even though that method worked well enough for experimenting, despite suboptimal performance, there is another obvious drawback: emscripten_sleep is implemented via setTimeout(), which can be very imprecise and is generally not reliable for fluid animation. Browsers actually have an API specifically for that use case: window.requestAnimationFrame(), but Taisei's original blocking control flow style is simply not compatible with it. Emscripten exposes this API with its emscripten_set_main_loop(), which the eventloop backend now uses on that platform. Unfortunately, C is still C, with no fancy closures or coroutines. With blocking calls into menu/scene loops gone, the control flow is reimplemented via so-called (pun intended) "call chains". That is basically an euphemism for callback hell. With manual memory management and zero type-safety. Not that the menu system wasn't shitty enough already. I'll just keep telling myself that this is all temporary and will be replaced with scripts in v1.4. * improve build system for emscripten + various fixes * squish menu bugs * improve emscripten event loop; disable EMULATE_FUNCTION_POINTER_CASTS Note that stock freetype does not work without EMULATE_FUNCTION_POINTER_CASTS; use a patched version from the "emscripten" branch here: https://github.com/taisei-project/freetype2/tree/emscripten * Enable -Wcast-function-type Calling functions through incompatible pointers is nasal demons and doesn't work in WASM. * webgl: workaround a crash on some browsers * emscripten improvements: * Persist state (config, progress, replays, ...) in local IndexDB * Simpler HTML shell (temporary) * Enable more optimizations * fix build if validate_glsl=false * emscripten: improve asset packaging, with local cache Note that even though there are rules to build audio bundles, audio does *not* work yet. It looks like SDL2_mixer can not work without threads, which is a problem. Yet another reason to write an OpenAL backend - emscripten supports that natively. * emscripten: customize the html shell * emscripten: force "show log" checkbox unchecked initially * emscripten: remove quit shortcut from main menu (since there's no quit) * emscripten: log area fixes * emscripten/webgl: workaround for fullscreen viewport issue * emscripten: implement frameskip * emscripter: improve framerate limiter * align List to at least 8 bytes (shut up warnings) * fix non-emscripten builds * improve fullscreen handling, mainly for emscripten * Workaround to make audio work in chromium emscripten-core/emscripten#6511 * emscripten: better vsync handling; enable vsync & disable fxaa by default
2019-03-09 20:32:32 +01:00
if(m->context) {