Commit graph

8040 commits

Author SHA1 Message Date
wangqr
638f1eb99e Detect byte order
Since uchardet does not do that for us

Fix wangqr/Aegisub#38
2020-04-03 22:21:00 -04:00
wangqr
799d96602d Fix uchardet enable macro
It should be applied to libaegisub
2020-04-03 22:19:25 -04:00
wangqr
01f6e86901 Modernize CMake (part 2) 2020-03-31 02:26:31 -04:00
wangqr
c0c2a9100d Modernize CMake (part 1) 2020-03-30 20:32:16 -04:00
wangqr
5318838b93 Migrate tests/setup.bat to PowerShell
Removed set / ping hacks
2020-03-29 22:25:15 -04:00
wangqr
62708e8e5c Ensure stdcall on XAudio2 callback
Fix for x86 build
2020-03-29 22:18:49 -04:00
wangqr
4b096e6c6e Merge branch 'dev-small-changes' into dev 2020-03-20 20:15:18 -04:00
wangqr
ef99ce14fd Detect iconv type at compile time instead of project generation 2020-03-20 08:03:44 -04:00
wangqr
88c486aca4 Change default audio output to XAudio2 on Windows
Downmix defaults to disabled on non-Windows OSes, because we don't support non-mono output at the moment
2020-03-20 04:13:41 -04:00
wangqr
3351235434 Move macOS specific build issue to Wiki 2020-03-14 01:00:44 -04:00
wangqr
64d0f86d8d Add git_version.h generation to CMake 2020-03-11 17:07:33 -04:00
wangqr
3394f22a93 Expand layer SpinCtrl a little
to compensate wxMac bug with GetSizeFromText
2020-03-10 01:16:28 -04:00
wangqr
f71919f2ae Fix missing config option for Align on macOS 2020-03-09 21:02:33 -04:00
wangqr
b6b21930fa Use homebrew addon instead of manually executing 2020-03-09 20:23:48 -04:00
wangqr
7cad0e06f9 Remove redundant email notification 2020-03-09 20:01:24 -04:00
wangqr
3096376ee6 Revert special naming for LuaJIT shared object
This was a hack on macOS when LuaJIT is dynamic only, and no longer needed as we are also using static LuaJIT on macOS now
2020-03-09 19:56:24 -04:00
wangqr
460cf9a9fd Drop support for macOS < 10.10
They are less than 2.4% in all macOS and 0.2% in all OS
2020-03-09 19:50:56 -04:00
wangqr
db855e8cf5 Fix Travis warning 2020-03-09 19:47:07 -04:00
wangqr
ed83c39807 Add back unit test to macOS
This reverts commit 272c2f48f1.
2020-03-09 17:36:52 -04:00
wangqr
52032dc14b Upgrade LuaJIT to 2.1.0-beta3 2020-03-09 13:01:09 -04:00
wangqr
3a50ba0386 Fix osx-fix-libs.py
Create symbolic links, to make libicu happy
Rewrite the script in python3, as python2 is deprecated
2020-03-08 19:17:53 -04:00
C.W. Betts
e7b64aa741 Poke the OS X icons.
More recent versions of OS X caused the old icons to display incorrectly/as garbage.
2020-03-08 17:48:52 -04:00
wangqr
8034fbbd22 Fix missing config option 2020-03-08 03:36:46 -04:00
wangqr
272c2f48f1 Ignore unit tests on macOS for now 2020-03-08 03:24:04 -04:00
wangqr
b557dfcb6e Update build instructions in README
CMake is not tested on macOS now, so don't include it in README. Be more specific on where to find dependencies.
2020-03-08 03:06:04 -04:00
wangqr
3556c9629e Remove IME handling code for Scintilla
The code was for wxCocoa, which is a dead implementation of wxWidgets. wxOSX/Cocoa does not need this hack anymore. And the code is causing linking errors due to using private structures in wxCocoa.

Revert fffb138b81
2020-03-08 01:48:58 -05:00
wangqr
f8bd53c479 Add macOS to travis 2020-03-08 01:37:27 -05:00
wangqr
50f63a54d0 Rewrite build instructions 2020-03-07 20:10:45 -05:00
wangqr
6cad7c3d6c Fix build issue with AviSynth on linux 2020-03-06 16:18:52 -05:00
wangqr
6bd3f4c26b Use target name without directory in $*_OBJ macro
Fix Aegisub/Aegisub#171
2020-02-17 15:08:29 +08:00
wangqr
cd7ee8d505 Use the first matched font in GDI font lister
Previously, when reading font data, we only set FaceName to the matched font. When we are using some font with special font weight (e.g. @Source Han Sans J Heavy), if we do not correct the font weight in the LOGFONTW struct, then subsequent call to get_font_data will fall back to default font. This causes wrongly matching Arial.ttf to any font that does not provide standard font weights.

Instead of only correcting FaceName using the matched font, we simply use the first matched font, thus the FaceName, Weight, CharSet, etc. will all be correct. This also eliminates the memcpy.
2020-01-05 17:12:02 -05:00
wangqr
cb77c0b395 Add option to downmix FFMS audio
When enabled, restore FFMS to old behavior, downmixing auduo to S16 mono. This can reduce cache usage.

Fix wangqr/Aegisub#31
2019-12-26 17:31:59 -05:00
wangqr
80ef2fbf35 Submit "align to video" on double click
Fix wangqr/Aegisub#34
2019-12-25 18:38:42 -05:00
wangqr
20fc9043e8 Don't shrink the size of layer SpinCtrl on GTK3 2019-11-21 23:04:51 -05:00
wangqr
a520f8a4a3 Rerender video display after EVT_PAINT
Otherwise the video display will be blank after restore from minimized.
2019-11-10 23:17:43 -05:00
wangqr
1bda6052b6 Use GetSizeFromText instead of GetSizeFromTextSize(GetTextExtent) 2019-11-07 16:06:53 -05:00
wangqr
557e81be1d VideoDisplay: Move render code to render callback
Previously, Render is called every time when the content is updated from event callbacks. So simply moving the mouse around will spam the event queue and cause video stuck. Now we do render in render callback, and only set a flag to indicate the need of re-render.
2019-11-04 07:03:45 -05:00
wangqr
f2676ddc3a Avisynth audio provider: add option to allow no downmix 2019-11-02 22:26:02 -04:00
Charlie Jiang
a2fc4bf479 🐛 Fix: lagi: Don't close iconv_t if it is iconv_invalid
When certain iconv_t is invalid, iconv_invalid(0xfffffffffffffffffffffff) will be returned. Passing it to iconv_close will cause a SIGSEGV.
2019-11-02 21:23:27 -04:00
wangqr
71894fd769 Fix bugs in ram audio provider
There was some magic bit operations to calculate the cache block offsets. This only works when both bytes_per_sample and channels are power of 2. Originally the format is assumed to be int16 mono, which satisfies this requirement. However in case we use original audio data, the channels can be something not a power of 2 (e.g. for 5.1 channel audio the number of channels is 6). This will break the calculation. We rewrite the calculation, without using those bit operations.
2019-11-02 02:26:36 -04:00
wangqr
56434c5f58 Fix bugs in separating Int16Mono and original audio
* A memory leak in FillBufferInt16Mono
* A now unnecessary assert
2019-11-02 00:41:14 -04:00
Luke Gorrie
45cd713ef9 lj_str.c: Remove special-case string interning fastpath
lj_str_new() had a separate fast-path and slow-path. This was bad
because (a) the fast-path was complex and (b) the fast-path was
actually slower than the slow-path in practice and (c) in practice it
could cause confusing performance problems depending on the memory
alignment of any often-reused string buffers in a program.

This change specifically makes the 'life' benchmark faster and more
robust to memory layout.
2019-11-02 00:23:50 -04:00
wangqr
29fd12258f Fix lint warnings about unexpected */
Introduced in b86238456f
2019-11-02 00:23:50 -04:00
wangqr
b947116937 Separate XAudio2 from DirectSound in preference and CMake 2019-11-02 00:23:50 -04:00
wangqr
309996aeb2 Remove the usage of __FILE__
Prepare for reproducible build
2019-11-02 00:23:44 -04:00
wangqr
2929b9db37 XAudio2: Try orignal sound format first, and fallback to 16bit mono on failure 2019-11-01 21:58:11 -04:00
wangqr
2ca9cf0de0 Fix unit tests
As we have removed the int16 convert and downmix related stuff.
2019-11-01 04:12:35 -04:00
wangqr
0336779735 Added experimental XAudio2 audio player
Removed downsampling in FFMS2 and CreateConvertAudioProvider, to ensure we can get the original audio channels and data.

Fix Aegisub/Aegisub#160
2019-11-01 03:32:42 -04:00
wangqr
b86238456f Change channels number and bytes per second to hardcoded value
Because these audio players can only handle 16bit mono anyway.
2019-10-30 01:16:28 -04:00
wangqr
620033915a Handle channel > 1 in hd & ram cache audio provider 2019-10-30 00:47:45 -04:00