June 2025 - Wine-gitlab - list.winehq.org

Re: [PATCH v2 0/6] MR8408: server: Use event syncs for more objects. - approved
by Alexandre Julliard (＠julliard) 24 Jun '25

24 Jun '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8408

1 0

[PATCH 0/1] MR8422: find/tests: Fix out-of-bound access to input in mangle_text.
by Yuxuan Shui (＠yshui) 24 Jun '25

24 Jun '25

input is not null-terminated so we need to check for length first. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8422

2 1

[PATCH 0/1] MR8420: dsound: Make sure to null-terminate strings for callbacks.
by Yuxuan Shui (＠yshui) 24 Jun '25

24 Jun '25

WideCharToMultiByte normally null-terminates the output if input length is -1, but it doesn't do so if the output buffer is too small. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8420

3 2

[PATCH v30 0/9] MR7650: Add support for AF_UNIX sockets (followup)
by Ralf Habacker (＠rhabacker) 24 Jun '25

24 Jun '25

Follow-up of !2786, which appears to have been abandoned. Depends on !8182 # Execute minimal test case 1. in dlls/ws2_32/tests/sock.c wrap all calls to test functions after `Init()` until `test_afunix()` with #if 0 ... #endif 2. checkout wine into ~/src/wine 3. build into ~/src/wine-build 4. run `(cd ~/src/wine-build; WINEPREFIX=~/src/wine/.wine ../wine/tools/runtest -q -P wine -T . -M ws2_32.dll -p dlls/ws2_32/tests/i386-windows/ws2_32_test.exe sock)` # How to debug wineserver Since `wineserver` runs in the background, simple calls to printf() will show nothing when the test case is executed. A workaround is to open a file and call fprintf() to write debug messages to this file, which can then be inspected. -- v30: ws2_32: Add note in bind() for AF_UNIX sockets ws2_32/tests: In tests for AF_UNIX sockets print actual directory used ws2_32/tests: In AF_UNIX socket tests always print used path in case of errors ws2_32/tests: Add test for AF_UNIX sockets server: Fix getsockname() and accept() on AF_UNIX sockets. server: Introduce error when attempting to create a SOCK_DGRAM AF_UNIX socket. server: Allow for deletion of socket files. ws2_32: Add support for AF_UNIX sockets. ws2_32: Add afunix.h header. https://gitlab.winehq.org/wine/wine/-/merge_requests/7650

3 9

[PATCH v2 0/1] MR8416: windows.storage: Split ApplicationData
by Ignacy Kuchciński 24 Jun '25

24 Jun '25

Windows registry keys such as `HKEY_LOCAL_MACHINE\\Software\\Microsoft\\WindowsRuntime\\ActivatableClassId\\Windows.Storage.ApplicationData` indicate that `Windows.Storage.ApplicationData` related classes should be in `windows.storage.applicationdata.dll`, and the remaining classes together with `Windows.Storage.StorageFile` should be in `windows.storage.dll`. This allows future implementions of class `Windows.Storage.StorageFile` to be correctly put in the `windows.storage.dll`, instead of `windows.storage.applicationdata.dll`, see https://gitlab.winehq.org/wine/wine/-/merge_requests/8394 -- v2: windows.storage: Split ApplicationData https://gitlab.winehq.org/wine/wine/-/merge_requests/8416

1 1

[PATCH v2 0/1] MR8418: d3dx9/tests: Fix volume test boxes.
by Yuxuan Shui (＠yshui) 24 Jun '25

24 Jun '25

This test seems to be testing different filters when creating volumes, and nothing indicates it's intentionally testing volume creation from out-of-bound srcboxes. The pixels array is 16x4 bytes, pixel format D3DFMT_A8R8G8B8 is 4 bpp, so pixels have 16 pixels in total. The boxes have 2 slices, with a slice pitch of 32 bytes, that's 8 pixel/slice. The original width*height is 6*6 which is thus beyond these bounds. This commit changed it to 4*2 so the rowpitch of 16 bytes is perserved. -- v2: d3dx9/tests: Fix volume test boxes. https://gitlab.winehq.org/wine/wine/-/merge_requests/8418

2 1

[PATCH 0/1] MR8416: windows.storage: Split ApplicationData
by Ignacy Kuchciński 24 Jun '25

24 Jun '25

Windows registry keys such as `HKEY_LOCAL_MACHINE\\Software\\Microsoft\\WindowsRuntime\\ActivatableClassId\\Windows.Storage.ApplicationData` indicate that `Windows.Storage.ApplicationData` related classes should be in `windows.storage.applicationdata.dll`, and the remaining classes together with `Windows.Storage.StorageFile` should be in `windows.storage.dll`. Split Windows.Storage.ApplicationData definitions out of windows.storage.idl to windows.storage.applicationdata.idl and import that instead of whole window.storage.idl into windows.storage.applicationdata.dll, to better reflect Windows implementation. This allows future implementions of class `Windows.Storage.StorageFile` to be correctly put in the `windows.storage.dll`, instead of `windows.storage.applicationdata.dll`, see https://gitlab.winehq.org/wine/wine/-/merge_requests/8394 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8416

3 3

[PATCH 0/6] MR8415: mf: Get rid of the IMFMediaShutdownNotify interface.
by Rémi Bernon 24 Jun '25

24 Jun '25

As discussed in !8402. @cmccarthy what do you think of this? I think queueing an event right before shutting down the queue is enough to get the media session notified? It won't get the actual event but it will be able to see the MF_E_SHUTDOWN error after calling `EndGetEvent`? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8415

2 8

[PATCH v5 0/2] MR8258: Draft: comctl32/treeview: Return from TREEVIEW_LButtonDown when the treeview handle is invalid.
by Jacob Czekalla (＠JacobCzekalla) 24 Jun '25

24 Jun '25

comctl32/treeview: Return from TREEVIEW_LButtonDown when the treeview handle is invalid. LButtonDown should return when the treeview handle is invalid (e.g. destroyed) after NM_CLICK to prevent further message processing. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=58211 -- v5: comctl32/treeview: Return from TREEVIEW_LButtonDown when the treeview handle is invalid. comctl32/tests: Add a test for treeview deletion during NM_CLICK in LBUTTONDOWN. https://gitlab.winehq.org/wine/wine/-/merge_requests/8258

4 6

[PATCH 0/2] MR8331: urlmon: Fix out-of-bound access when trying to remove dot segments from URI.
by Yuxuan Shui (＠yshui) 24 Jun '25

24 Jun '25

Dot segments removal in parse_canonicalize makes the assumption that the entire path part of the URI fits into the output buffer (len is the length the output URI would have taken had the output buffer been large enough). So when the output buffer is too small, remove_dot_segments will access out-of-bound. Considering URI unescaping could be happening at the same time, there is no simple way to calculate the required output buffer size correctly without keeping the actual output URI in memory. * * * As you can see from the [test results](https://testbot.winehq.org/JobDetails.pl?Key=158742), this seems to be bugged on native. I don't know if we need to replicate native's behavior -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8331

3 6