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

[PATCH 0/5] MR7205: winecrt0: Add builtin functions for msvc/arm mode
by Martin Storsjö 28 Feb '25

28 Feb '25

Since 9b8409fce4da18e3d9a914a9e5831d10eb9052de, a PE compiler is required for 32 bit arm. That can either be supplied by using plain Clang (from a distro), or llvm-mingw. However when using plain Clang (in MSVC mode), we're lacking compiler builtin functions that either would be provided by MSVC libraries or by compiler-rt libraries bundled in llvm-mingw. Vendor a copy of the relevant files from compiler-rt and include them when building for arm in MSVC mode. This allows building a functional wine for 32 bit arm without requiring third party tools such as llvm-mingw. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7205

3 22

[PATCH 0/6] MR7236: winewayland: Clipboard support
by Alexandros Frantzis (＠afrantzis) 28 Feb '25

28 Feb '25

This MR adds clipboard support to the winewayland driver. Under Wayland only applications that have the keyboard focus can interact with the clipboard (a.k.a. `wl_data_device`). Such constraints are not a natural fit for Wine's current clipboard infrastructure, which uses a separate thread and window in the desktop process to act as the win32-side clipboard proxy for all native windows. This MR tries to work within the current Wine clipboard Wine by forwarding relevant clipboard messages to the foreground window which is likely to have the keyboard focus and that can actually handle them. This works well in practice (although there are some edge cases this fails), but I am open to different ideas about implementing the clipboard integration. Some notes about the MR: 1. Some formats that require special treatment (e.g., CF_HTML, CF_HDROP) are not implemented in this MR to keep the size reasonable (and also to not distract from the goal of this MR which is to propose/discuss the basic design of clipboard integration). 2. Dynamic registration of newly-seen/unknown formats is not supported at the moment. It's not clear to what degree that's useful, since many (most?) Windows clipboard format strings are free-form and don't use the MIME type standard. But perhaps enough Windows apps know about MIME types nowadays to make this addition worth it? 3. Since access to the Wayland clipboard data is performed from the focused window threads (rather than the dedicated clipboard thread), any blocking while waiting for data to be sent/received has more potential to affect the applications. I have implemented reasonable timeouts which are hopefully enough to make this a non-issue. A (more complex) alternative would be a different design that attempts to offload any potentially blocking work to the dedicated clipboard thread (or is there some kind of async read/write mechanism we can use?). I am not convinced the extra complexity is worth it, though. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7236

8 28

[PATCH 0/1] MR6977: apisetschema: Add some Windows 10 API sets.
by Tim Clem (＠tclem) 26 Feb '25

26 Feb '25

Needed for Indiana Jones and the Great Circle. --- The naming of the shellfolders one is not a mistake; they seemed to have named it differently than all the other shell API sets. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6977

3 4

[PATCH 0/1] MR4662: properly track refcount and dependencies with forwarded exports.
by YUAN RUI (＠number201724) 25 Feb '25

25 Feb '25

fix: https://bugs.winehq.org/show_bug.cgi?id=52094 tests: ``` #include <Windows.h> #ifdef FROM_IMPORT #pragma comment(lib, "python3.lib") extern "C" int Py_IsInitialized(); #endif int main() { #ifdef FROM_IMPORT // test from import table Py_IsInitialized( ); #else // //Test ref count int (*Py_IsInitialized)(); HMODULE hmod_310 = LoadLibraryA( "C:\\Python310\\python310.dll" ); HMODULE hmod = LoadLibraryA( "C:\\Python310\\python3.dll" ); if ( !hmod ) return 0; *(FARPROC*)&Py_IsInitialized = GetProcAddress( hmod, "Py_IsInitialized" ); FreeLibrary( hmod_310 ); Py_IsInitialized( ); FreeLibrary( hmod ); // test dependencies hmod = LoadLibraryA( "C:\\Python310\\python3.dll" ); if ( !hmod ) return 0; *(FARPROC*)&Py_IsInitialized = GetProcAddress( hmod, "Py_IsInitialized" ); Py_IsInitialized(); FreeLibrary( hmod ); #endif return 0; } ``` -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4662

5 5

[PATCH 0/3] MR3124: Move periodic registry save out of wineserver
by Paul Gofman (＠gofman) 25 Feb '25

25 Feb '25

The problem is that registry save is a very heavy operation (scheduled each 30sec in wineserver) during which server doesn't process any requests and the whole prefix is stalled for the duration of the operation. For some reference, the process takes from 50-100ms here up to 1-1.5sec with default initial registry (after some registry modifications which trigger actual registry flush), depending on the filesystem type and state (as huge time may be spent in file close / rename). With the same registry after this patchset the server part (flush_key returning the whole registry data) is taking ~4-5mcs, measured from the client side so that already includes data transfer. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3124

5 9

[PATCH v9 0/1] MR3096: ntdll: Add NtGetNextProcess stub.
by Alex Henrie (＠alexhenrie) 25 Feb '25

25 Feb '25

The stub is enough for TokenViewer.exe to start up and display its user interface. Based on a patch by Nikolay Sivov. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=45119 -- v9: ntdll: Add NtGetNextProcess stub. https://gitlab.winehq.org/wine/wine/-/merge_requests/3096

2 3

[PATCH 0/3] MR7182: winex11: Keep track and use expected window focus time on WM_STATE changes.
by Rémi Bernon 19 Feb '25

19 Feb '25

This is an attempt at improving X11 focus handling, and will most importantly be useful to later better handle dynamic window decoration changes on any mutter-based window manager. These window managers do a little dance in that case, unmapping/mapping the window every time decoration are added/removed, and we often get confused by the focus events it triggers. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7182

2 6

[PATCH 0/2] MR7142: win32u: Reset surface clip region if it matches the visible rect.
by Rémi Bernon 18 Feb '25

18 Feb '25

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=57665 With an empty clipping region, when the window is being shown, the Expose event we receive is ignored by `expose_window_surface`, assuming there's no redraw necessary. If a clipping region is set, we call `NtUserRedrawWindow` with `RDW_ERASE` which clears the window. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7142

1 3

[PATCH v3 0/1] MR7179: wined3d: Add some DXT format mappings in the Vulkan renderer.
by Aida Jonikienė 18 Feb '25

18 Feb '25

This removes some texture errors when using Himavant. -- v3: wined3d: Add DXT format mappings in the Vulkan renderer. https://gitlab.winehq.org/wine/wine/-/merge_requests/7179

3 4

[PATCH 0/5] MR4894: setupapi: implement CM_Locate_DevNode
by Etaash Mathamsetty (＠etaash.mathamsetty) 17 Feb '25

17 Feb '25

This api is used by libxess on intel GPUs. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4894

3 8