September 2025 - Wine-gitlab

[PATCH v5 0/11] MR8995: Draft: ntdll/kernelbase: Add accurate node count feature to prevent memory access violations initializing multiple physical nodes as a single logical one on multi-node systems
by Danny Waser (＠wasertech) 19 Sep '25

19 Sep '25

A much cleaner result (written by hand) that we can reuse where needed later (_i.e._ `GetNumaProcessorNode` which also crashes with [a similar MAV](https://gist.github.com/wasertech/f894ce8d6250e72a01a861c0e4eb6064) on multi-node systems). Took me a while to understand where I should put everything, not even sure I got it right. Let me know. I thought I was going to need `FileNumaNodeInformation` but turns out it's not really needed. I can remove it if you want. This is what I see when I try to get a node count read: ```log 0024:fixme:ntdll:init_numa_info node affinity; using node 0. GetNumaHighestNodeNumber: 1 ``` I only see this fixme if I try to access `FILE_NUMA_NODE_INFORMATION`. I would really prefer such a behavior for our compatibility layer on any multi-node system. A big thanks to @besentv and @zfigura for their invaluable feedback on this. -- v5: ntdll: simplify return handling from `init_numa_info` in `NtQuerySystemInformationEx( SystemNumaProcessorMap )` ntdll: refactor init_numa_info to return NTSTATUS and improve error handling ntdll: remove unused definition https://gitlab.winehq.org/wine/wine/-/merge_requests/8995

2 11

[PATCH 0/1] MR9007: wmic: Add a basic help option "/?".
by Louis Lenders (＠xe) 19 Sep '25

19 Sep '25

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=45343 The installers from that bugreport use this (in Win7 mode) to check if wmic is present. (The installer still fails after this patch as findstr.exe lacks some support: "match_regexp character ranges (i.e. [abc], [^a-z]) are not supported") -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9007

3 2

[PATCH 0/1] MR8998: vbscript: Allow redim without a prior dim.
by Elizabeth Figura (＠zfigura) 19 Sep '25

19 Sep '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/8998

3 3

[PATCH 0/4] MR9005: ShowWindow paramater usage
by Alistair Leslie-Hughes (＠alesliehughes) 19 Sep '25

19 Sep '25

This patchset is just to correct the parameter passed to ShowWindow. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9005

4 6

[PATCH 0/3] MR9006: Several fixes & improvements when winedbg is started in auto mode.
by eric pouech (＠epo) 19 Sep '25

19 Sep '25

This series covers various fixes & improvements when a crash happens in a program which fires up winedbg in auto mode. Firstly, winedbg is a bit too strict for detecting the startup sequence (and fails to display the backtrace in these stricter cases). Third patch relaxes the startup conditions in that case. Secondly, if failing programs is a GUI, winedbg is launched in a new console (since GUI are now detached from console), but console is closed after winedbg terminates (which is after printing the crash details). So rather not useful :-( To work around this, conhost in its configuration dialog has tab 'Configuration' / 'End of program' / 'Close console' checkbox But which is not implemented. Second patch add supports for this in conhost. Lastly, conhost has the ability to store configuration per application. Hence, one can configure conhost with specific settings for winedbg. Interesting options are: - setting very high buffer zone so that conhost keeps track of all winedbg output, - unchecking the 'Close console' option. To set it for winedbg, - start it with './wine wineconsole winedbg' - use 'Properties' from the menu (right-click in conhost) to set the options for winedbg Since 68f3a8e699904bd308de0da4a1e1d2c9eb91ea96, conhost doesn't pick the registry entry for winedbg (if it exists) when started with './wine wineconsole winedbg'. Third patch solves that in a somehow hacky way, but couldn't find a better option that doesn't break https://bugs.winehq.org/show_bug.cgi?id=10941 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9006

2 3

[PATCH 0/2] MR9003: makedep: Add support for generating winmd files.
by Hans Leidekker (＠hans) 19 Sep '25

19 Sep '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/9003

4 9

[PATCH 0/2] MR9004: win32u: Avoid swapping / flushing drawables with destroyed windows.
by Rémi Bernon 19 Sep '25

19 Sep '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/9004

1 2

[PATCH v3 0/5] MR9000: Draft: win32u: Check the queue access time from the shared memory.
by Rémi Bernon 19 Sep '25

19 Sep '25

This is a step in the direction of continuously polling the queue fd on the server side, removing the need for ntsync to notify wineserver before / after waiting on the queue. For this my plan is to make win32u also notify wineserver to update the queue access time before waiting (through set_queue_mask), if it's been long enough. Then decide that a queue is hung based on whether it is signaled but access time is more than 5s ago (ie: neither get_message nor set_queue_mask are called regularly even though queue is signaled). On the server side, the queue fd would stop being polling once it gets ready, and win32u would notify wineserver after all the driver events have been processed, to start polling it again, as a separate request from the wait. -- v3: win32u: Check the queue access time from the shared memory. server: Use monotonic_time as queue access time base. win32u: Move the message queue access time to shared memory. server: Move hooks_count to the end of the queue_shm_t struct. server: Return early if there's no queue in queue requests. https://gitlab.winehq.org/wine/wine/-/merge_requests/9000

1 6

[PATCH v2 0/5] MR9000: Draft: win32u: Check the queue access time from the shared memory.
by Rémi Bernon 19 Sep '25

19 Sep '25

This is a step in the direction of continuously polling the queue fd on the server side, removing the need for ntsync to notify wineserver before / after waiting on the queue. For this my plan is to make win32u also notify wineserver to update the queue access time before waiting (through set_queue_mask), if it's been long enough. Then decide that a queue is hung based on whether it is signaled but access time is more than 5s ago (ie: neither get_message nor set_queue_mask are called regularly even though queue is signaled). On the server side, the queue fd would stop being polling once it gets ready, and win32u would notify wineserver after all the driver events have been processed, to start polling it again, as a separate request from the wait. -- v2: win32u: Check the queue access time from the shared memory. server: Use monotonic_time as queue access time base. win32u: Move the message queue access time to shared memory. server: Move hooks_count to the end of the queue_shm_t struct. server: Return early if there's no queue in queue requests. https://gitlab.winehq.org/wine/wine/-/merge_requests/9000

1 6

Re: [PATCH v6 0/2] MR8983: wineserver: Request RLIMIT_NOFILE maximum allowed value.
by Alexandre Julliard (＠julliard) 19 Sep '25

19 Sep '25

Ah right, I guess I got confused by the macOS version numbering, sorry about that. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8983#note_116343

1 0