--
v6: winex11: Use EM_REPLACESEL to set IME composition strings.
include: Add some immdev.h definitions.
winex11: Always zero terminate XIM composition string.
winex11: Compute preedit text buffer sizes in WCHAR units.
winex11: Use an IME UI message to set composition status.
imm32: Only send GCS_CURSORPOS when cursor position changes.
winex11: Use an IME UI message to set cursor pos.
winex11: Use an IME UI message to get cursor pos.
winex11: Use an IME UI message to set open status.
win32u: Keep the builtin IME UI window in win32u struct imc.
imm32: Add a struct ime field to track whether IME is builtin.
https://gitlab.winehq.org/wine/wine/-/merge_requests/2637
The tests shows that, for sink writer created by MFCreateSinkWriterFromURL,
we are not able to grab the media sink object immediately after the creation,
unless the media sink object can be created without setting media type(MP3, ASF).
--
v3: mfreadwrite/tests: Test GetService for MP4 sink writer.
mfreadwrite/tests: Add more tests for MFCreateSinkWriterFromURL.
https://gitlab.winehq.org/wine/wine/-/merge_requests/2690
According to the tests in 23b72ad, when we are reading a compressed stream, the type returned by `stream_props_GetMediaType()` should reflect compressed format even if we finnally output uncomressed data. For example, if we use wmvcore reader to read a WMV3 stream and output RGB24, the format information returned by `stream_props_GetMediaType()` should be WMV3, not RGB24.
--
v7: winegstreamer: Use stream format in stream_props_GetMediaType.
winegstreamer: Implement amt_from_wg_format_video_wmv.
winegstreamer: Implement wg_parser_stream_get_codec_format.
winegstreamer: Introduce format_is_compressed.
https://gitlab.winehq.org/wine/wine/-/merge_requests/2387
According to the tests in 23b72ad, when we are reading a compressed stream, the type returned by `stream_props_GetMediaType()` should reflect compressed format even if we finnally output uncomressed data. For example, if we use wmvcore reader to read a WMV3 stream and output RGB24, the format information returned by `stream_props_GetMediaType()` should be WMV3, not RGB24.
This patch set is marked as draft now, because PATCH 2 and PATCH 3 is affected by 3e8936a2 in MR !2258. So I'll submit a newer version of this after !2258 get merged.
--
v6: winegstreamer: Use stream format in stream_props_GetMediaType.
winegstreamer: Implement amt_from_wg_format_video_wmv.
winegstreamer: Implement wg_parser_stream_get_stream_format.
winegstreamer: Introduce format_is_compressed.
https://gitlab.winehq.org/wine/wine/-/merge_requests/2387
The tests shows that, for sink writer created by MFCreateSinkWriterFromURL,
we are not able to grab the media sink object immediately after the creation,
unless the media sink object can be created without setting media type(MP3, ASF).
--
v2: mfreadwrite/tests: Test getting media sink immediately after MFCreateSinkWriterFromURL.
mfreadwrite/tests: Add more tests for MFCreateSinkWriterFromURL.
https://gitlab.winehq.org/wine/wine/-/merge_requests/2690
Unpolished. Raises framerate by around 15% in SotTR because vk_sets_mutex
is locked only once for each descriptor heap. It's possible to collate
multiple writes for a single vkUpdateDescriptorSets() call as was done
in the optimised copying path, but it may not be worth it.
A possible complication with this: if it becomes possible to remove the
descriptor mutexes after the remaining problems with queue sequencing are
fixed, we theoretically may still need them for this. Descriptors should
not be updated while binding, but there's no reason more can't be written
in another thread while a command list is submitted to a queue.
--
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/71