February 2026 - Wine-gitlab

[PATCH 0/1] MR8712: strmbase: Get rid of stream_cs.
by Conor McCarthy (＠cmccarthy) 02 Feb '26

02 Feb '26

The following sequence can occur: sink_EndOfStream, lock stream_cs, renderer sink_eos, strmbase_passthrough_eos, IMediaSeeking_GetStopPosition, MediaSeekingPassThru_GetStopPosition, get_connected, IPin_ConnectedTo, lock filter_cs And at the same time: filter_Stop, lock filter_cs, dmo_wrapper_cleanup_stream, lock stream_cs Resulting in deadlock. This can occur in Space Engineers. dmo_wrapper_cleanup_stream() locks stream_cs to access `pin.peer` and `pin.pAllocator`, which are normally protected by filter_cs, but this lock was added to fix a test failure, so I don't know if there's an alternative. I'm concerned about the possible performance impact of removing stream_cs though. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8712

4 3

[PATCH v2 0/4] MR9935: winegstreamer: Fix some small issues with the resampler.
by Conor McCarthy (＠cmccarthy) 02 Feb '26

02 Feb '26

These fixes will be needed for the topology loader patches. -- v2: mf/tests: Test resampler output type after setting the input type. https://gitlab.winehq.org/wine/wine/-/merge_requests/9935

2 6

[PATCH 0/12] MR10009: mf/topology_loader: Rewrite the topology loader to match more closely the test results.
by Conor McCarthy (＠cmccarthy) 01 Feb '26

01 Feb '26

The end result is more a rewrite than a modified version of the original. I was able to split the changes into patches, but this is not necessarily easier to understand than doing a mostly complete replacement in one commit and then adding a few things on top of that. With the change to video converter creation, mfmediaengine now depends on ENABLE_XVP being set, as it does in Windows. This change has already been made in mfmediaengine. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10009

2 12

[PATCH 0/2] MR9996: msvcrt: Correct backslash handling in argument parser.
by Yeshun Ye (＠yeyeshun) 01 Feb '26

01 Feb '26

The `cmdline_to_argv` function previously had a bug in handling backslashes preceding double quotes, particularly regarding the escaping rules. \ When processing a case like `"\"c:\windows\\""`, the flawed logic incorrectly consumed the quote character, leading to an error in parsing the number of arguments. Signed-off-by: YeshunYe yeyeshun(a)uniontech.com -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9996

2 4

[PATCH v14 0/4] MR8394: Fix Clip Studio Paint crash with Windows 10 compatibility mode
by Ignacy Kuchciński (＠ignapk) 01 Feb '26

01 Feb '26

This fixes the crash when launching ClipStudioPaint.exe with Windows 10 compatiblity mode by adding the minimum required windows.ui and windows.storage definitions and stubs. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=58254 -- v14: windows.storage: Add stubs for StorageFile class. include: Add Windows.Storage.StorageFile missing definitions. windows.ui: Add stubs for RadialController classes. include: Add Windows.UI.Input.RadialController definitions. https://gitlab.winehq.org/wine/wine/-/merge_requests/8394

2 5

[PATCH v5 0/3] MR10007: Support enabling/disabling IME on winewayland
by Ziia Shi (＠mkrsym1) 01 Feb '26

01 Feb '26

Currently, IME is always enabled for all applications. This can cause some very annoying issues on winewayland. For example, fcitx takes over the character repeat functionality when enabled, and does it via repeatedly sending sequences of keyup/keydown, which leads to applications not being able to track hold duration of a button (can be reproduced on any game which requires holding a letter key). The spec states that the enable request should only be sent when a text edit is focused, and a disable request should be sent when it is unfocused. Windows provides a "similar-ish" mechanism: applications can use ImmAssociateContext/ImmAssociateContextEx to associate an IMC when IME should be enabled, and set associated IMC to null when it should be disabled. From my rather limited testing, it is implemented this way in Chromium/CEF, QT and Unity. Setting it to null internally results in deactivation of the currently associated context. This MR makes the ImeSetActiveContext function of the builtin IME utilize a new driver function to communicate activation status. It does not take into account which specific context was activated, thus semi-stub. This fixes behavior of applications using the above-mentioned components. I only implemented it for winewayland, it probably should also be implemented for winex11 too, but I'm not sure how. ~~Unfortunately this breaks IME input in Wine comctl32 components. I am not sure how it is supposed to behave on Windows or how to test it at the current point, will look into it later. Having IME on by default appears to "fix" it, but that is very likely not the Windows behavior; moreover, at least Unity does not disable IME on start, which causes the above-mentioned issue to occur until a textbox is manually focused and unfocused.~~ -- v5: winewayland: Implement SetIMEEnabled driver call. https://gitlab.winehq.org/wine/wine/-/merge_requests/10007

2 4

[PATCH v13 0/4] MR8394: Fix Clip Studio Paint crash with Windows 10 compatibility mode
by Ignacy Kuchciński (＠ignapk) 01 Feb '26

01 Feb '26

This fixes the crash when launching ClipStudioPaint.exe with Windows 10 compatiblity mode by adding the minimum required windows.ui and windows.storage definitions and stubs. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=58254 -- v13: windows.storage: Add stubs for StorageFile class. include: Add Windows.Storage.StorageFile missing definitions. windows.ui: Add stubs for RadialController classes. include: Add Windows.UI.Input.RadialController definitions. https://gitlab.winehq.org/wine/wine/-/merge_requests/8394

2 6

[PATCH v4 0/3] MR10007: Draft: Support enabling/disabling IME on winewayland
by Ziia Shi (＠mkrsym1) 01 Feb '26

01 Feb '26

Currently, IME is always enabled for all applications. This can cause some very annoying issues on winewayland. For example, fcitx takes over the character repeat functionality when enabled, and does it via repeatedly sending sequences of keyup/keydown, which leads to applications not being able to track hold duration of a button (can be reproduced on any game which requires holding a letter key). The spec states that the enable request should only be sent when a text edit is focused, and a disable request should be sent when it is unfocused. Windows provides a "similar-ish" mechanism: applications can use ImmAssociateContext/ImmAssociateContextEx to associate an IMC when IME should be enabled, and set associated IMC to null when it should be disabled. From my rather limited testing, it is implemented this way in Chromium/CEF, QT and Unity. Setting it to null internally results in deactivation of the currently associated context. This MR makes the ImeSetActiveContext function of the builtin IME utilize a new driver function to communicate activation status. It does not take into account which specific context was activated, thus semi-stub. This fixes behavior of applications using the above-mentioned components. I only implemented it for winewayland, it probably should also be implemented for winex11 too, but I'm not sure how. Unfortunately this breaks IME input in Wine comctl32 components. I am not sure how it is supposed to behave on Windows or how to test it at the current point, will look into it later. Having IME on by default appears to "fix" it, but that is very likely not the Windows behavior; moreover, at least Unity does not disable IME on start, which causes the above-mentioned issue to occur until a textbox is manually focused and unfocused. -- v4: winewayland: Implement SetIMEEnabled driver call. https://gitlab.winehq.org/wine/wine/-/merge_requests/10007

2 4

[PATCH v2 0/3] MR10007: Draft: Support enabling/disabling IME on winewayland
by Ziia Shi (＠mkrsym1) 01 Feb '26

01 Feb '26

Currently, IME is always enabled for all applications. This can cause some very annoying issues on winewayland. For example, fcitx takes over the character repeat functionality when enabled, and does it via repeatedly sending sequences of keyup/keydown, which leads to applications not being able to track hold duration of a button (can be reproduced on any game which requires holding a letter key). The spec states that the enable request should only be sent when a text edit is focused, and a disable request should be sent when it is unfocused. Windows provides a "similar-ish" mechanism: applications can use ImmAssociateContext/ImmAssociateContextEx to associate an IMC when IME should be enabled, and set associated IMC to null when it should be disabled. From my rather limited testing, it is implemented this way in Chromium/CEF, QT and Unity. Setting it to null internally results in deactivation of the currently associated context. This MR makes the ImeSetActiveContext function of the builtin IME utilize a new driver function to communicate activation status. It does not take into account which specific context was activated, thus semi-stub. This fixes behavior of applications using the above-mentioned components. I only implemented it for winewayland, it probably should also be implemented for winex11 too, but I'm not sure how. Unfortunately this breaks IME input in Wine comctl32 components. I am not sure how it is supposed to behave on Windows or how to test it at the current point, will look into it later. Having IME on by default appears to "fix" it, but that is very likely not the Windows behavior; moreover, at least Unity does not disable IME on start, which causes the above-mentioned issue to occur until a textbox is manually focused and unfocused. -- v2: winewayland: Implement SetIMEEnabled driver call. https://gitlab.winehq.org/wine/wine/-/merge_requests/10007

2 3

[PATCH 0/3] MR10007: Draft: Support enabling/disabling IME on winewayland
by Ziia Shi (＠mkrsym1) 01 Feb '26

01 Feb '26

Currently, IME is always enabled for all applications. This can cause some very annoying issues on winewayland. For example, fcitx takes over the character repeat functionality when enabled, and does it via repeatedly sending sequences of keyup/keydown, which leads to applications not being able to track hold duration of a button (can be reproduced on any game which requires holding a letter key). The spec states that the enable request should only be sent when a text edit is focused, and a disable request should be sent when it is unfocused. Windows provides a "similar-ish" mechanism: applications can use ImmAssociateContext/ImmAssociateContextEx to associate an IMC when IME should be enabled, and set associated IMC to null when it should be disabled. From my rather limited testing, it is implemented this way in Chromium/CEF, QT and Unity. Setting it to null internally results in deactivation of the currently associated context. This MR makes the ImeSetActiveContext function of the builtin IME utilize a new driver function to communicate activation status. It does not take into account which specific context was activated, thus semi-stub. This fixes behavior of applications using the above-mentioned components. I only implemented it for winewayland, it probably should also be implemented for winex11 too, but I'm not sure how. Unfortunately this breaks IME input in Wine comctl32 components. I am not sure how it is supposed to behave on Windows or how to test it at the current point, will look into it later. Having IME on by default appears to "fix" it, but that is very likely not the Windows behavior; moreover, at least Unity does not disable IME on start, which causes the above-mentioned issue to occur until a textbox is manually focused and unfocused. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10007

2 3