March 2024 - wine-gitlab

[PATCH v2 0/2] MR5323: win32u: Key auto-repeat improvements.
by Alexandros Frantzis (＠afrantzis) 15 Mar '24

15 Mar '24

A couple of win32u auto-repeat related improvements: 1. Cancel auto-repeat only if the repeat key is released (not other keys). 2. Cancel previous key auto-repeat when starting a new one for a different window (see commit message for more info). I have kept the changes minimal, which can lead to a few extraneous `kill_system_timer` calls. To improve on this would require more extensive tracking/clearing of `thread_info->key_repeat_msg`. -- v2: win32u: Cancel previous key auto-repeat when starting a new one. https://gitlab.winehq.org/wine/wine/-/merge_requests/5323

3 3

[PATCH v3 0/1] MR5188: Draft: winemac.drv: Hide app's dock icon when it wouldn't have a taskbar item on Windows.
by Tim Clem (＠tclem) 15 Mar '24

15 Mar '24

This is a change from the default behavior in macdrv for as long as it's existed, as far as I can tell. Previously, gaining a dock icon was a one-way path. However, that doesn't jive with the way taskbar entries work on Windows; if an app has no windows, it doesn't appear in the taskbar. This patch attempts to remedy cases where an app winds up with superfluous dock icons for exe's that no longer have windows (looking at you, basically every launcher and Steam). The major concern I can see with this is that if an app closes all of its windows but does not exit, there will be no indication that it is still running. Two thoughts on that: 1. That *should* be an anomalous case, such as the app hanging on exit. 2. Effectively the same behavior would happen on Windows. I would love to hear any other thoughts about this change. I'm open (though I would not prefer it) to defaulting the registry key to false if that would alleviate any concerns. -- v3: winemac.drv: Hide app's dock icon when it wouldn't have a taskbar item on Windows. https://gitlab.winehq.org/wine/wine/-/merge_requests/5188

4 5

[PATCH v5 0/2] MR714: vkd3d-shader/dxil: Implement DX intrinsic BufferUpdateCounter.
by Conor McCarthy (＠cmccarthy) 15 Mar '24

15 Mar '24

With no compiler support, there's no easy way to verify shader runner support for d3d11 UAV counters works, so I've left it out for now. -- v5: vkd3d-shader/dxil: Implement DX intrinsic BufferUpdateCounter. tests/hlsl: Add UAV counter tests. https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/714

4 5

[PATCH 0/5] MR5322: server: Clear queue QS_RAWINPUT if get_rawinput_buffer got everything.
by Rémi Bernon (＠rbernon) 15 Mar '24

15 Mar '24

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/5322

2 6

[PATCH 0/7] MR724: vkd3d-shader/hlsl: Improve SM1 support for non-float operations, part 4.
by Francisco Casas (＠fcasas) 15 Mar '24

15 Mar '24

SM1 support for LOGIC_NOT, LOGIC_OR, and LOGIC_AND. -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/724

4 12

[PATCH 0/4] MR5321: Minidump slowdowns (Part III)
by eric pouech (＠epo) 15 Mar '24

15 Mar '24

This is the last part for handling the minidump slow downs: - it removes the integration of RUNTIME_INFO chunks for x86_64 - adds fallback mechanism to read this unwind info from winedbg side. Some remarks: - it looks a bit awkward that SymFunctionTableAccess reads PE information from PE image on dbghelp side, while the unwind information reading is read (in this MR) from debugger side - some testings with native show that: + unwind info (inside StackWalk calls) is read through the read memory handler (so delegated to debugger) + there's no fallback from the read memory handler - so this looks like the minimal things to do - however, there could be additional fallback in dbghelp (like in dbghelp internal read memory handler); didn't spend time on these as they would only be helpful for non live targets, using the default read memory handler. - fixed winedbg to reload properly minidumps. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5321

2 4

[PATCH 0/1] MR723: vkd3d-shader/spirv: Implement MAD in two operations if flagged as precise.
by Conor McCarthy (＠cmccarthy) 15 Mar '24

15 Mar '24

This would eliminate the todo for the precise mad() test in !718. Maybe we need test results on nvidia and intel to decide if we actually want this. -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/723

3 3

[PATCH v3 0/5] MR5294: mfplat: Implement better VIDEOINFOHEADER2 support.
by Rémi Bernon (＠rbernon) 15 Mar '24

15 Mar '24

This is the first set of patches aimed at improving the functions to convert between media types and various representations. Not everything ended up being necessary for what I wanted to do (MFVIDEOFORMAT seems to be the only format able to carry aperture information after all) but the tests and implementation are now there and will still be useful to convert media types from/to AM_MEDIA_TYPE formats. -- v3: mfplat/mediatype: Implement MFInitAMMediaTypeFromMFMediaType for FORMAT_VideoInfo2. mfplat/tests: Add tests for MFInitMediaTypeFromVideoInfoHeader2. mfplat/mediatype: Implement MFInitMediaTypeFromVideoInfoHeader2. mfplat/mediatype: Implement implicit MFInitMediaTypeFromVideoInfoHeader subtype. mfplat/tests: Test each VIDEOINFOHEADER field conversion separately. https://gitlab.winehq.org/wine/wine/-/merge_requests/5294

4 7

[PATCH v4 0/2] MR711: vkd3d-shader/ir: Introduce a better CFG structurizer, part 6.
by Giovanni Mascellani (＠giomasce) 15 Mar '24

15 Mar '24

~~This applies on top of !704, the last two commits belong here.~~ Here we finally emit the structured program, replacing the older simple structurizer. The advantage of the new structurizer is that the CFG if statically described in the generated program, without having to trace the value of a variable. Upcoming MRs will make the CFG even simpler, easier to read and hopefully to optimize for downstream compilers. -- v4: vkd3d-shader/ir: Emit multilevel jumps in the structured program. vkd3d-shader/ir: Emit the reconstructed structured program. https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/711

3 6

[PATCH 0/1] MR5320: ntdll/tests: Fix a test failure.
by Hans Leidekker (＠hans) 15 Mar '24

15 Mar '24

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/5320

2 1

2025

2024

2023

2022

wine-gitlab March 2024