December 2025 - Wine-GitLab

[PATCH v8 0/6] MR7466: mf: Fix some issues with seeking.
by Conor McCarthy (＠cmccarthy) Feb. 2, 2026

Feb. 2, 2026

-- v8: mf: Initialise the grabber sample count when setting state after a seek. mf: Do not count draining a downstream node as a request fulfillment. mf: Do not count placing the stream sink ENDOFSEGMENT marker as a request fulfillment. mf/tests: Wait for session closure at the end of test_media_session_Start(). mf/tests: Test pause followed by immediate restart at current time in test_media_session_Start(). mf/tests: Test sample delivery where applicable in test_media_session_Start(). https://gitlab.winehq.org/wine/wine/-/merge_requests/7466

3 7

[PATCH 0/1] MR8712: strmbase: Get rid of stream_cs.
by Conor McCarthy (＠cmccarthy) Feb. 2, 2026

Feb. 2, 2026

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 v13 0/1] MR9061: cabinet: Add LZX compression support to FCI
by Eric Lasota (＠ejlasota) Jan. 31, 2026

Jan. 31, 2026

This adds LZX compression support to FCI. Most of the work for this was done here: https://github.com/elasota/liblzx The LZX compression code is based on the LZX compression code from wimlib with modifications to support the cabinet variant and streaming compression. Eric Biggers (the author of wimlib) has given permission to redistribute the modified LZX code under LGPLv2 terms: https://wimlib.net/forums/viewtopic.php?t=854 I've been testing this against some large data sets (dumped game assets, Visual Studio installation) with a test program and decompressing with 7-zip and diffs have been coming back OK. -- v13: cabinet: lzx compression support https://gitlab.winehq.org/wine/wine/-/merge_requests/9061

4 3

[PATCH v4 0/2] MR9639: Fix the integration of PowerPoint add-in in MS Office 2016
by Ivan Lyugaev (＠valy) Jan. 28, 2026

Jan. 28, 2026

Encountered an issue with PowerPoint add-in integration. Reviewing the logs revealed the problem was due to the lack of CLSID lookup via the CurVer section and incorrect handling of the PARAMFLAG_NONE flag. A test was conducted in Windows to verify the handling of PARAMFLAG_NONE, PARAMFLAG_FIN, and PARAMFLAG_FOUT. Test class: ``` interface ITest : IDispatch { [id(1)] HRESULT TestMethod(BSTR* value); [id(2)] HRESULT TestMethodOut([out] BSTR* retVal); [id(3)] HRESULT TestMethodIn([in] BSTR* inVal); }; ``` Results in Windows: ``` --------TestMethod-------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethod] Hello from VT_BSTR Invoke returned: 0x00000000 -------TestMethodIn------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethodIn] Hello from VT_BSTR Invoke returned: 0x00000000 -------TestMethodOut------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethodOut] (null) Invoke returned: 0x00000000 ``` Results in Wine: ``` --------TestMethod-------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethod] (null) Invoke returned: 0x00000000 -------TestMethodIn------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethodIn] Hello from VT_BSTR Invoke returned: 0x00000000 -------TestMethodOut------- Variant type: 0x0008 (VT_BSTR) [CTest::TestMethodOut] (null) ``` Therefore, parameters with the PARAMFLAG_NONE flag are implied to be input parameters in Windows and are handled accordingly as PARAMFLAG_FIN. Added tests for the implemented changes. In the last two tests (their code is provided below), the CLSID is nullified in the guid_from_string function, which is why this test is left with a todo_wine tag. ``` clsid = CLSID_StdFont; hr = CLSIDFromString(L"MyApp.DocumentTest.3", &clsid); todo_wine ok(IsEqualCLSID(&clsid, &CLSID_StdFont), "Unexpected clsid %s.\n", wine_dbgstr_guid(&clsid)); clsid = CLSID_StdFont; hr = CLSIDFromString(L"MyApp.DocumentTest.5", &clsid); todo_wine ok(IsEqualCLSID(&clsid, &CLSID_StdFont), "Unexpected clsid %s.\n", wine_dbgstr_guid(&clsid)); ``` -- v4: combase: Implement CLSID lookup with CurVer fallback oleaut32: Fix parameter value loss in ITypeInfo_Invoke for byref parameters https://gitlab.winehq.org/wine/wine/-/merge_requests/9639

3 8

[PATCH v16 0/7] MR9371: kernel32: Implement TH32CS_SNAPMODULE32 support for CreateToolhelp32Snapshot.
by Yuxuan Shui (＠yshui) Jan. 27, 2026

Jan. 27, 2026

Intention is to use `CreateToolhelp32Snapshot` in `dbghelp.dll: EnumerateLoadedModulesW64`. Right now we can't because `CreateToolhelp32Snapshot` can't capture 32bit modules in wow64 processes, which is needed to support `SYMOPT_INCLUDE_32BIT_MODULES`. The advantage of using `CreateToolhelp32Snapshot` in mainly performance. Right now `EnumerateLoadedModulesW64` is `O(n^2)`, because of each module it calls `GetModuleInformation` which is itself `O(n)` w.r.t. number of modules. Whereas with `CreateToolhelp32Snapshot` there is no need for `GetModuleInformation`, thus reduce the complexity to just `O(n)`. * * * P.S. I am also uncomfortable with the amount of similar code between `CreateToolhelp32Snapshot` and `EnumProcessModulesEx`. Can something be done here? -- v16: kernel32: Implement TH32CS_SNAPMODULE32 support for CreateToolhelp32Snapshot. kernel32: Use GetCurrentProcess() handle in CreateToolhelp32Snapshot if possible. kernel32/tests: Test CreateToolhelp32Snapshot with TH32CS_SNAPMODULE32. kernel32: Fix CreateToolhelp32Snapshot on old WoW64. include: Add TH32CS_SNAPMODULE32. kernel32/tests: Handle ERROR_BAD_LENGTH from CreateToolhelp32Snapshot. kernel32/tests: Fix CreateToolhelp32Snapshot failure check. https://gitlab.winehq.org/wine/wine/-/merge_requests/9371

4 10

[PATCH 0/1] MR8586: mshtml: Fix node's onclick not responding.
by Yongjie Yao (＠yaoyongjie) Jan. 26, 2026

Jan. 26, 2026

When installing the photoshop-setup.exe version of Photoshop 2025, there is no response after clicking the continue button. The reasons are as follows: As a property of the node, setting events such as onclick on the node does not respond when clicked. The following jscript code was executed during the installation of Photoshop 2025: $("#continueButton").attr("onclick", "continueBtnClicked('installOptionsContinueClicked'); return false;"); -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8586

4 11

[PATCH v4 0/3] MR7659: windows.devices.bluetooth: Add implementation for IBluetoothStatics::GetDefaultAsync(), BluetoothAdapter.
by Vibhav Pant (＠vibhavp) Jan. 26, 2026

Jan. 26, 2026

This MR adds a basic implementation for the BluetoothAdapter interface in Windows.Devices.Bluetooth, including winebth.sys support for getting properties associated with the radio's Low Energy capabilities. -- v4: windows.devices.bluetooth: Implement BluetoothAdapter::get_BluetoothAddress. windows.devices.bluetooth: Implement IBluetoothAdapterStatics::GetDefaultAsync(). windows.devices.bluetooth/tests: Add tests for IBluetoothAdapter. https://gitlab.winehq.org/wine/wine/-/merge_requests/7659

3 5

[PATCH 0/1] MR9486: windowscodecs: correct GIF LZW compression initialization
by Jiajin Cui (＠jin-king1) Jan. 22, 2026

Jan. 22, 2026

The GIF LZW compression algorithm was incorrectly initializing the code size to 8 bits regardless of the actual color depth. This caused issues with images that have fewer colors requiring smaller initial code sizes. The fix calculates the proper initial code size based on the actual number of colors in the image, using the formula: minimum code size = max(2, ceil(log2(color_count))). This ensures proper compression for images with small color palettes and maintains compatibility with GIF decoders that expect correct LZW initialization parameters. Log: Fixed GIF compression for images with small color palettes Signed-off-by: Jiajin Cui <cuijiajin(a)uniontech.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9486

6 9

[PATCH v13 0/3] MR7539: combase : clsid_from_string_reg() add read of CurVer key value.
by Maotong Zhang (＠xiaotong) Jan. 21, 2026

Jan. 21, 2026

If the read ProgID only has a CurVer key and no CLSID key, it will directly return CO_E_CCLASSSTRING, causing the program to fail to continue running. Add the key value for finding CurVer. -- v13: combase:clsid_from_string_reg() add read of CurVer key value. ole32:Do not zero out when not a CLSID ole32/tests:Remove now succeeding todo_wine https://gitlab.winehq.org/wine/wine/-/merge_requests/7539

3 5

[PATCH v7 0/5] MR9572: opengl32: Implement GL_EXT_memory_object_win32 for base handle types.
by Derek Lesho (＠dlesho) Jan. 20, 2026

Jan. 20, 2026

Not sure about the extension replacement mechanism. Also, should tests be failing when extensions aren't supported? -- v7: opengl32: Expose GL_EXT_memory_object_win32. win32u/opengl: Implement glImportMemoryWin32(Handle/Name)EXT for opaque handle types. opengl32, win32u: Move gl_error to wgl_context. https://gitlab.winehq.org/wine/wine/-/merge_requests/9572

3 6