taisei/meson.build

353 lines
11 KiB
Meson
Raw Normal View History

project('taisei', 'c',
license : 'MIT',
version : 'v1.3-dev',
meson_version : '>=0.45.0',
default_options : [
'c_std=c11',
2019-03-17 06:06:47 +01:00
'default_library=static',
'libzip:enable_crypto=false',
'libzip:enable_bzip2=false',
2019-03-20 08:35:29 +01:00
'sdl2:use_haptic=disabled',
'sdl2:use_power=disabled',
'sdl2:use_render=disabled',
'sdl2:use_sensor=disabled',
2019-03-18 07:05:10 +01:00
# You may want to change these for a debug build dir
'buildtype=release',
'strip=true',
'b_lto=true',
'b_ndebug=if-release',
]
)
minimum_recommended_meson_version = '0.48.0'
if meson.version().version_compare('<@0@'.format(minimum_recommended_meson_version))
warning('Old Meson version detected. Try upgrading to at least @0@ if the build fails.'.format(minimum_recommended_meson_version))
endif
cc = meson.get_compiler('c')
python3 = import('python3').find_python()
macos_app_bundle = get_option('macos_bundle') and host_machine.system() == 'darwin'
subdir('scripts')
config = configuration_data()
taisei_c_args = [
'-Wall',
'-Wpedantic',
'-Werror=assume',
'-Werror=implicit-function-declaration',
'-Werror=incompatible-pointer-types',
#
# Keep the rest sorted
#
'-Wabsolute-value',
'-Wcast-align',
'-Wcast-align=strict',
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
'-Wcast-function-type',
'-Wclobbered',
'-Wduplicated-branches',
'-Wduplicated-cond',
'-Wfloat-overflow-conversion',
'-Wfloat-zero-conversion',
'-Wfor-loop-analysis',
'-Wformat-pedantic',
'-Wformat-security',
'-Wgcc-compat',
'-Wgnu',
'-Wignored-qualifiers',
'-Wimplicit-fallthrough',
'-Winit-self',
'-Wlogical-op',
'-Wmissing-prototypes',
'-Wno-long-long',
'-Wno-missing-braces',
2019-02-08 20:12:52 +01:00
'-Wno-typedef-redefinition',
'-Wnull-dereference',
'-Wparentheses',
'-Wshadow=compatible-local',
'-Wsometimes-uninitialized',
'-Wstrict-prototypes',
'-Wtype-limits',
Some compatibility fixes for Elbrus compiler (and others based on EDG front end) (#157) * Reduce alignment to 16 due to stack variables alignment On some archs (e.g. elbrus) there is no possibility to align variables on stack by more that 16 bytes. * Added check for -Wno-typedef-redefinition In C11, there is legal to redefine typedef with the same type, but some compilers like elbrus's lcc and, probably, intel's icc generate a warning about that. So we disable it. * Typo fix * Fix __builtin_unreachable() warning on some frontends like EDG * Added check for __attribute__((designated_init)) It doesn't work for elbrus'c lcc and clang; but clang does not generate warning by default, but we still disable it just in case. * A hack to avoid -Wtype-limits for compilers with unsigned enums * Rewritten unreachable __builtin_unreachable() behavior. According to https://bugs.llvm.org/show_bug.cgi?id=13910, old clang versions also suffer from this issue along with lcc. And also this warning option is removed from gcc since 4.5.0, and it's likely there is no unreachable code analysis at all: https://gcc.gnu.org/onlinedocs/gcc-4.5.0/gcc/Warning-Options.html * Fixed unneeded stray UNREACHABLE which caused warnings * Removed alignas() which is unneeded after lowering alignment requirements * Removed a hack of -Wunreachable-code disabling on certain compilers Actually, there was a wrong opinion that most of unreachable code warnings are produced by __builtin_unreachable() itself, not the certain use of it. After commit 36493d3, it is finally figured out that it wasn't the case. * Added dummy element to a struct in a test for designated_init This test will enexpectedly fail when GNU extensions are not available
2019-02-07 09:11:13 +01:00
'-Wunneeded-internal-declaration',
2019-02-08 20:12:52 +01:00
'-Wunreachable-code',
'-Wunreachable-code-loop-increment',
]
if get_option('b_pch') and meson.version().version_compare('<0.50.0')
# Workaround for Meson bug: https://github.com/mesonbuild/meson/issues/4905
taisei_c_args += ['-fpch-deps']
endif
taisei_c_args = cc.get_supported_arguments(taisei_c_args)
foreach arglist : [
['-msse', '-mfpmath=sse'],
]
if cc.has_multi_arguments(arglist)
taisei_c_args += arglist
endif
endforeach
sm_check = run_command(check_submodules_command)
if sm_check.stdout() != ''
foreach line : sm_check.stdout().strip().split('\n')
warning(line)
endforeach
endif
if sm_check.stderr() != ''
warning('Submodule check completed with errors:\n@0@'.format(sm_check.stderr()))
endif
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 = get_option('static') or (host_machine.system() == 'emscripten')
2019-03-17 20:05:32 +01:00
dep_freetype = dependency('freetype2', required : true, static : static, fallback : ['freetype', 'freetype_dep'])
dep_png = dependency('libpng', version : '>=1.5', required : true, static : static, fallback : ['libpng', 'png_dep'])
2019-03-18 07:05:10 +01:00
dep_sdl2 = dependency('sdl2', version : '>=2.0.5', required : true, static : static, fallback : ['sdl2', 'sdl2_dep'])
dep_sdl2_mixer = dependency('SDL2_mixer', required : false, static : static, fallback : ['sdl2_mixer', 'sdl2_mixer_dep'])
2019-03-17 21:58:15 +01:00
dep_webp = dependency('libwebp', version : '>=0.5', required : false, static : static, fallback : ['libwebp', 'webp_dep'])
dep_webpdecoder = dependency('libwebpdecoder', version : '>=0.5', required : false, static : static, fallback : ['libwebp', 'webpdecoder_dep'])
dep_zlib = dependency('zlib', required : true, static : static, fallback : ['zlib', 'zlib_dep'])
2019-03-17 06:06:47 +01:00
dep_zip = dependency('libzip', version : '>=1.2', required : false, static : static, fallback : ['libzip', 'libzip_dep'])
dep_crypto = dependency('libcrypto', required : false, static : static)
dep_m = cc.find_library('m', required : false)
dep_cglm = subproject('cglm').get_variable('cglm_dep')
dep_glad = subproject('glad').get_variable('glad_dep')
taisei_deps = [
dep_cglm,
dep_freetype,
dep_m,
dep_png,
dep_sdl2,
dep_zlib,
# don't add glad here
]
2018-10-19 01:51:59 +02:00
if dep_webpdecoder.found()
taisei_deps += dep_webpdecoder
elif dep_webp.found()
warning('libwebpdecoder not found, will link against libwebp instead')
taisei_deps += dep_webp
else
error('libwebpdecoder or libwebp is required, neither found')
endif
if host_machine.system() == 'windows'
taisei_deps += cc.find_library('shlwapi')
endif
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 host_machine.system() == 'emscripten'
package_data = false
enable_zip = false
else
package_data = get_option('package_data')
enable_zip = get_option('enable_zip')
package_data = (package_data == 'auto' ? enable_zip : package_data == 'true')
endif
if enable_zip
2019-03-17 06:06:47 +01:00
assert(dep_zip.found(), 'ZIP support enabled but libzip not found')
taisei_deps += dep_zip
endif
if package_data and not enable_zip
error('ZIP support must be enabled for data packaging to work')
endif
config.set('TAISEI_BUILDCONF_USE_ZIP', taisei_deps.contains(dep_zip))
have_posix = cc.has_header_symbol('unistd.h', '_POSIX_VERSION')
have_vla = not cc.has_header_symbol('stdlib.h', '__STDC_NO_VLA__')
have_complex = not cc.has_header_symbol('stdlib.h', '__STDC_NO_COMPLEX__')
have_timespec = cc.has_header_symbol('time.h', 'timespec_get')
assert(have_vla and have_complex, 'Your C implementation needs to support complex numbers and variable-length arrays.')
config.set('TAISEI_BUILDCONF_HAVE_TIMESPEC', have_timespec)
config.set('TAISEI_BUILDCONF_HAVE_INT128', cc.sizeof('__int128') == 16)
config.set('TAISEI_BUILDCONF_HAVE_LONG_DOUBLE', cc.sizeof('long double') > 8)
config.set('TAISEI_BUILDCONF_HAVE_POSIX', have_posix)
config.set('TAISEI_BUILDCONF_HAVE_MAX_ALIGN_T', cc.compiles('#include <stddef.h>\nmax_align_t i;', name : 'max_align_t test'))
prefer_relpath_systems = [
'windows',
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
]
force_relpath_systems = [
'emscripten',
]
if macos_app_bundle
bundle_dir = 'Taisei.app'
datadir = join_paths(bundle_dir, 'Contents', 'Resources')
bindir = join_paths(bundle_dir, 'Contents', 'MacOS')
config.set('TAISEI_BUILDCONF_RELATIVE_DATA_PATH', true)
# arguments must be strings...
meson.add_install_script(
python3.path(),
join_paths(meson.source_root(), 'scripts', 'macos-install-dylibs.py'),
':'.join(meson.get_cross_property('macos_lib_path', [])),
':'.join(meson.get_cross_property('macos_tool_path', [])),
meson.get_cross_property('macos_tool_prefix', ''),
)
else
datadir = get_option('datadir')
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 force_relpath_systems.contains(host_machine.system())
config.set('TAISEI_BUILDCONF_RELATIVE_DATA_PATH', true)
elif get_option('install_relative') == 'auto'
config.set('TAISEI_BUILDCONF_RELATIVE_DATA_PATH', prefer_relpath_systems.contains(host_machine.system()))
else
config.set('TAISEI_BUILDCONF_RELATIVE_DATA_PATH', get_option('install_relative') == 'true')
endif
2017-12-21 05:52:25 +01:00
if config.get('TAISEI_BUILDCONF_RELATIVE_DATA_PATH')
OpenGL ES 3.0 rendering backend (#148) * First steps towards shader transpilation Needs to be manually enabled via -Dshader_transpiler=true. Requires shaderc. https://github.com/google/shaderc Not yet functional due to missing SPIRV-Cross integration. SPIRV-Cross currently does not have an official C API, and crossc is too minimal to be useful. The current plan is to extend crossc and vendor it, while also sending PRs upstream. * Integrate crossc; shader transpilation for GLES now works * fix leak * gles30 backend now playable on Mesa with 3.2 context Some rendering issues are present. Identified so far: - Marisa's lasers are invisible - Death effect looks wrong Also, a small pixmap manipulation library has been written, and the texture uploading API redesigned around it. * fix marisa lasers in GLES (uniform name clashed with builtin) * fix player death effect in GLES (another name clash) * Dump ANGLE's translated shader code in debug log * fix screenshots * Drop support for triangle fans, switch to strips Fans offer no advantage over strips, and they've been removed in D3D10+, so ANGLE has to emulate them. * crude workaround for an ANGLE bug * Re-enable GL debug labels, fix an issue with them that affected ANGLE (but was always technically a bug) * fix race condition in shaderc initialization * New SDL_RWops interface for vertex buffers * Optimize VBO streaming via buffering updates Measurable performance improvement even with the main gl33 renderer, drastic improvement with ANGLE. * Fix the depth texture binding problem under ANGLE Apparently it hates GL_DEPTH_COMPONENT16 for some reason. Sized internal formats are not supported in GLES 2.0 anyway, so not using them is probably a good idea. * fix GLES2.0 segfault (the backend still doesn't work, though) * dump GL extensions at info log level, not debug * get around a Mesa bug; more correct texture format table for GLES2 * Correct GLES3 texture format table according to the spec Not a Mesa bug after all * require crossc>=1.5.0, fallback to subproject * Request at least 8bit per color channel in GL backends * Forbid lto for static windows builds with shader_transpiler=true * fix edge case segfault * Add basic ANGLE bundling support to the build system Windows only, and no NSIS support yet * Fix various windows-related build system and installer brokenness * Disable gles backends by default * update documentation
2018-10-02 00:36:10 +02:00
bindir = '.'
2017-12-21 05:52:25 +01:00
else
bindir = get_option('bindir')
endif
endif
if get_option('install_freedesktop') == 'auto'
install_xdg = not config.get('TAISEI_BUILDCONF_RELATIVE_DATA_PATH')
else
install_xdg = get_option('install_freedesktop') == 'true'
endif
if config.get('TAISEI_BUILDCONF_RELATIVE_DATA_PATH')
data_path = 'data'
doc_path = '.' # Meson bug https://github.com/mesonbuild/meson/issues/4295
xdg_path = 'freedesktop.org'
2017-12-21 05:52:25 +01:00
# This is relative to SDL_GetBasePath()
config.set_quoted('TAISEI_BUILDCONF_DATA_PATH', data_path)
2017-12-21 05:52:25 +01:00
if macos_app_bundle
# Actual installation path
data_path = join_paths(datadir, data_path)
# I don't know why would you do that, but more power to you
xdg_path = join_paths(datadir, xdg_path)
endif
else
data_path = join_paths(datadir, 'taisei')
config.set_quoted('TAISEI_BUILDCONF_DATA_PATH', join_paths(get_option('prefix'), data_path))
doc_path = join_paths(datadir, 'doc', 'taisei')
2017-12-21 04:49:44 +01:00
xdg_path = datadir
endif
is_debug_build = get_option('buildtype').startswith('debug')
is_developer_build = (get_option('developer') == 'auto' ? is_debug_build : get_option('developer') == 'true')
config.set('TAISEI_BUILDCONF_DEVELOPER', is_developer_build)
config.set('TAISEI_BUILDCONF_LOG_FATAL_MSGBOX', (
host_machine.system() == 'windows' or
host_machine.system() == 'darwin' or
not is_developer_build
))
config.set('TAISEI_BUILDCONF_DEBUG_OPENGL', get_option('debug_opengl'))
install_docs = get_option('docs') and host_machine.system() != 'emscripten'
if host_machine.system() == 'windows'
if install_docs
custom_target('COPYING.txt',
command : [eolconv_command, host_eol_style, '@INPUT@', '@OUTPUT@'],
input : 'COPYING',
output : 'COPYING.txt',
install : true,
install_dir : doc_path
)
endif
OpenGL ES 3.0 rendering backend (#148) * First steps towards shader transpilation Needs to be manually enabled via -Dshader_transpiler=true. Requires shaderc. https://github.com/google/shaderc Not yet functional due to missing SPIRV-Cross integration. SPIRV-Cross currently does not have an official C API, and crossc is too minimal to be useful. The current plan is to extend crossc and vendor it, while also sending PRs upstream. * Integrate crossc; shader transpilation for GLES now works * fix leak * gles30 backend now playable on Mesa with 3.2 context Some rendering issues are present. Identified so far: - Marisa's lasers are invisible - Death effect looks wrong Also, a small pixmap manipulation library has been written, and the texture uploading API redesigned around it. * fix marisa lasers in GLES (uniform name clashed with builtin) * fix player death effect in GLES (another name clash) * Dump ANGLE's translated shader code in debug log * fix screenshots * Drop support for triangle fans, switch to strips Fans offer no advantage over strips, and they've been removed in D3D10+, so ANGLE has to emulate them. * crude workaround for an ANGLE bug * Re-enable GL debug labels, fix an issue with them that affected ANGLE (but was always technically a bug) * fix race condition in shaderc initialization * New SDL_RWops interface for vertex buffers * Optimize VBO streaming via buffering updates Measurable performance improvement even with the main gl33 renderer, drastic improvement with ANGLE. * Fix the depth texture binding problem under ANGLE Apparently it hates GL_DEPTH_COMPONENT16 for some reason. Sized internal formats are not supported in GLES 2.0 anyway, so not using them is probably a good idea. * fix GLES2.0 segfault (the backend still doesn't work, though) * dump GL extensions at info log level, not debug * get around a Mesa bug; more correct texture format table for GLES2 * Correct GLES3 texture format table according to the spec Not a Mesa bug after all * require crossc>=1.5.0, fallback to subproject * Request at least 8bit per color channel in GL backends * Forbid lto for static windows builds with shader_transpiler=true * fix edge case segfault * Add basic ANGLE bundling support to the build system Windows only, and no NSIS support yet * Fix various windows-related build system and installer brokenness * Disable gles backends by default * update documentation
2018-10-02 00:36:10 +02:00
if angle_enabled
install_data(
get_option('angle_libgles'),
get_option('angle_libegl'),
install_dir : join_paths(bindir, 'ANGLE'),
)
endif
else
if install_docs
install_data('COPYING', install_dir : doc_path)
endif
OpenGL ES 3.0 rendering backend (#148) * First steps towards shader transpilation Needs to be manually enabled via -Dshader_transpiler=true. Requires shaderc. https://github.com/google/shaderc Not yet functional due to missing SPIRV-Cross integration. SPIRV-Cross currently does not have an official C API, and crossc is too minimal to be useful. The current plan is to extend crossc and vendor it, while also sending PRs upstream. * Integrate crossc; shader transpilation for GLES now works * fix leak * gles30 backend now playable on Mesa with 3.2 context Some rendering issues are present. Identified so far: - Marisa's lasers are invisible - Death effect looks wrong Also, a small pixmap manipulation library has been written, and the texture uploading API redesigned around it. * fix marisa lasers in GLES (uniform name clashed with builtin) * fix player death effect in GLES (another name clash) * Dump ANGLE's translated shader code in debug log * fix screenshots * Drop support for triangle fans, switch to strips Fans offer no advantage over strips, and they've been removed in D3D10+, so ANGLE has to emulate them. * crude workaround for an ANGLE bug * Re-enable GL debug labels, fix an issue with them that affected ANGLE (but was always technically a bug) * fix race condition in shaderc initialization * New SDL_RWops interface for vertex buffers * Optimize VBO streaming via buffering updates Measurable performance improvement even with the main gl33 renderer, drastic improvement with ANGLE. * Fix the depth texture binding problem under ANGLE Apparently it hates GL_DEPTH_COMPONENT16 for some reason. Sized internal formats are not supported in GLES 2.0 anyway, so not using them is probably a good idea. * fix GLES2.0 segfault (the backend still doesn't work, though) * dump GL extensions at info log level, not debug * get around a Mesa bug; more correct texture format table for GLES2 * Correct GLES3 texture format table according to the spec Not a Mesa bug after all * require crossc>=1.5.0, fallback to subproject * Request at least 8bit per color channel in GL backends * Forbid lto for static windows builds with shader_transpiler=true * fix edge case segfault * Add basic ANGLE bundling support to the build system Windows only, and no NSIS support yet * Fix various windows-related build system and installer brokenness * Disable gles backends by default * update documentation
2018-10-02 00:36:10 +02:00
if angle_enabled
error('install_angle is only available for Windows targets at the moment')
endif
endif
systype = (have_posix ? 'POSIX (@0@)' : '@0@').format(host_machine.system())
systype = '@0@, @1@, @2@'.format(systype, host_machine.cpu_family(), host_machine.cpu())
if meson.is_cross_build()
systype = '@0@ (cross-compiling)'.format(systype)
endif
version_deps = []
subdir('misc')
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
subdir('emscripten')
subdir('external')
subdir('resources')
subdir('doc')
subdir('xdg')
subdir('atlas')
subdir('src')
summary = '''
Summary:
Taisei version: @0@
System type: @1@
Build type: @2@
Audio backends: @3@ (default: @4@)
Renderers: @5@ (default: @6@)
Shader translation: @7@
ZIP support: @8@
Package data: @9@
Relative install paths: @10@
Prefix: @11@
Executables: @12@
Data: @13@
Documentation: @14@
'''.format(
taisei_version_string,
systype,
get_option('buildtype'),
', '.join(enabled_audio_backends),
get_option('a_default'),
', '.join(enabled_renderers),
get_option('r_default'),
get_option('shader_transpiler'),
enable_zip,
package_data,
config.get('TAISEI_BUILDCONF_RELATIVE_DATA_PATH'),
get_option('prefix'),
# the $ is intentional
join_paths('$prefix', bindir),
join_paths('$prefix', data_path),
join_paths('$prefix', doc_path)
)
message(summary)
run_command(postconf_command)