February 2026 - Wine-GitLab

Re: [PATCH v9 0/34] MR9928: Draft: dsound: Speed up resampling.
by Matteo Bruni (＠Mystral) Feb. 19, 2026

Feb. 19, 2026

I think I've got some test results strongly implying that Windows 10+ uses cubic interpolation, not a sinc / FIR filter, after all. I'll try to confirm that and, if that's right, I'll probably attempt to write an implementation from scratch, making use of the knowledge I gathered over time. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9928#note_129991

1 0

[PATCH v2 0/8] MR10134: ntdll: Fix various issues with mapping Unix timezone to Windows timezone.
by Paul Gofman (＠gofman) Feb. 19, 2026

Feb. 19, 2026

-- v2: ntdll: Accept 30min shift in match_tz_date(). ntdll: Better search for DST change in find_dst_change(). ntdll: Drop additional TZ matching with match_tz_name(). ntdll: Try to determine system time zone to Windows zone by name. ntdll: Factor out read_reg_tz_info(). ntdll: Fix get_timezone_info() for timezones defined with inverse DST on Unix. server: FIx timezone bias for timezones defined with inverse DST on Unix. https://gitlab.winehq.org/wine/wine/-/merge_requests/10134

2 10

[PATCH v2 0/1] MR10128: windowscodecs: Add fallback path to copypixels_to_24bppBGR().
by Dmitry Timoshkov (＠dmitry) Feb. 19, 2026

Feb. 19, 2026

-- v2: windowscodecs: Add fallback path to copypixels_to_24bppBGR(). https://gitlab.winehq.org/wine/wine/-/merge_requests/10128

3 3

[PATCH 0/8] MR10134: ntdll: Fix various issues with mapping Unix timezone to Windows timezone.
by Paul Gofman (＠gofman) Feb. 19, 2026

Feb. 19, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10134

2 10

[PATCH v3 0/3] MR10137: url: Implement InetIsOffline
by Si Lai (＠LesBoys43) Feb. 19, 2026

Feb. 19, 2026

Sending upstream -- v3: url: Add missing include https://gitlab.winehq.org/wine/wine/-/merge_requests/10137

2 5

[PATCH 0/1] MR10139: win32u: Only update host window state if clip_client changes.
by Rémi Bernon Feb. 19, 2026

Feb. 19, 2026

Otherwise we might race with other host window events coming in and attempting to change the win32 state as well. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10139

1 1

[PATCH v2 0/2] MR10137: url: Implement InetIsOffline
by Si Lai (＠LesBoys43) Feb. 19, 2026

Feb. 19, 2026

Sending upstream -- v2: url: Rebase https://gitlab.winehq.org/wine/wine/-/merge_requests/10137

3 4

[PATCH 0/1] MR10137: url: Implement InetIsOffline
by Si Lai (＠LesBoys43) Feb. 19, 2026

Feb. 19, 2026

Sending upstream -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10137

2 1

[PATCH 0/1] MR10136: joy.cpl: Avoid hangs when closing the panel.
by Brendan Shanks (＠bshanks) Feb. 19, 2026

Feb. 19, 2026

This may not be 100% reproducible but if you open joy.cpl and go to the XInput page, then the Windows.Gaming.Input page, then click Cancel, the dialog will not close and cannot be closed with either OK or Cancel. The issue is that PSN_KILLACTIVE is sent when navigating away from a page, and then PSN_RESET is sent when the dialog is closed (i.e. OK/Cancel) even if the page is not currently selected. For the sequence above, test_xi_dialog_proc() receives PSN_SETACTIVE, then PSN_KILLACTIVE, then PSN_RESET. On PSN_RESET, since the events and threads have already been closed, execution gets stuck in the MsgWaitForMultipleObjects() loop (so the dialog still works, but there's no way to exit). To fix this, ensure that the PSN_KILLACTIVE/PSN_RESET block only runs if the input thread is running. 9427e8a975199f70a0f029a690e1f622aff6d256 previously fixed this issue for the DInput page, where the UnregisterDeviceNotification() call led to a use-after-free and crash. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10136

3 2

[PATCH 0/1] MR10135: wow64win: SPI_GETDEFAULTINPUTLANG result is pointer-sized.
by Esme Povirk (＠madewokherd) Feb. 19, 2026

Feb. 19, 2026

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59404 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10135

2 1