April 2026 - Wine-GitLab - list.winehq.org

[PATCH v5 0/3] MR10743: small odbc fixes
by Ivan Ivlev (＠iviv) April 27, 2026

April 27, 2026

1)odbc32: compare strings case-insensitively in get_driver_filename() Driver names should be compared without case-sensetivity. unixODBC has no problems when testDSN got "Driver = SQLite3" instead of "Driver = SQLITE3" https://learn.microsoft.com/en-us/openspecs/sql_server_protocols/ms-odbcstr… 2)odbc32: skip set_value for NULL filename in replicate_odbc_to_registry() get_driver_filename can return NULL, for example when \~/.odbc.ini contains driver that has been deleted. We want to prevent SIGSEGV because of wcslen(filename) when filename is NULL. 3)odbc32: check if row_number is NULL in param_options_unix() According to documentation, SQLParamOptions can accept NULL row_number. Now when NULL row_nuber is passed, crash happens 4)odbc32: make sure ODBC_CALL returns SQLRETURN Not sure about this one, need opinion. The targeted problem: now ODBC_CALL return SQLRETURN which is truncated from NTSTATUS. If WINE_UNIX_CALL fails at some point, ODBC_CALL returns truncated NTSTATUS error code, STATUS_ACCESS_VIOLATION (0xC0000005) truncated to 5, then function gets invalid SQLRETURN. The following solution checks of NTSTATUS is actually short SQLRETURN casted to NTSTATUS. -- v5: odbc32: Check if row_number is NULL in param_options_unix(). odbc32: Skip operations for NULL filename in replicate_odbc_to_registry(). https://gitlab.winehq.org/wine/wine/-/merge_requests/10743

3 6

[PATCH v6 0/4] MR10104: user32/edit: set bCaptureState flag after SetCapture()
by Ivan Ivlev (＠iviv) April 27, 2026

April 27, 2026

Edit control does not release mouse capture after WM_LBUTTONDOWN has been processed two times in a row. This MR fixes this issue by releasing capture and setting it again on WM_LBUTTONDOWN if window was already captured. I have stumbled upon the app in which edit control process WM_LBUTTONDOWN twice on one mouse click. While in windows it causes no problem, wine edit does not release capture, so other buttons inside this app become unclickable. It happens because when WM_LBUTTONDOWN is processed twice, WM_CAPTURECHANGED is sent and bCaptureState is set to FALSE, then EDIT_WM_LButtonUp does not release capture because bCaptureState is already FALSE. Thing about tests: I have added test case for this problem to be visible. I also had to use todo_wine because SendMessageA for WM_LBUTTONDOWN and WM_LBUTTONUP returns 0 in wine and 1 in windows (which is not the case for combobox test, where SendMessageА for WM_LBUTTONDOWN also returns 1 in wine). -- v6: comctl32_v6/edit: Set bCaptureState flag after SetCapture(). user32/edit: Set bCaptureState flag after SetCapture(). comctl32/tests/edit.c: Test WM_LBUTTONDOWN processing. user32/tests/edit.c: Test WM_LBUTTONDOWN processing. https://gitlab.winehq.org/wine/wine/-/merge_requests/10104

2 5

[PATCH v3 0/1] MR10757: vbscript: Fix call depth decrement on early returns from exec_script.
by Francis De Brabandere (＠francisdb) April 27, 2026

April 27, 2026

The early-return paths in exec_script skip the call_depth decrement that the normal exit path performs, so any of them leak the depth counter and can spuriously trigger VBSE_OUT_OF_STACK. This is a fix for a regression introduced in !10592 -- v3: vbscript: Move call_depth check past exec_script's setup phase. https://gitlab.winehq.org/wine/wine/-/merge_requests/10757

2 1

[PATCH v2 0/1] MR10757: vbscript: Fix call depth decrement on early returns from exec_script.
by Francis De Brabandere (＠francisdb) April 27, 2026

April 27, 2026

The early-return paths in exec_script skip the call_depth decrement that the normal exit path performs, so any of them leak the depth counter and can spuriously trigger VBSE_OUT_OF_STACK. This is a fix for a regression introduced in !10592 -- v2: vbscript: Decrement call depth on early returns from exec_script. https://gitlab.winehq.org/wine/wine/-/merge_requests/10757

2 2

[PATCH v2 0/3] MR10710: wineandroid: prepare ioctl dispatch for Java-only thread execution (part of !10569)
by Twaik Yont (＠twaik) April 27, 2026

April 27, 2026

Part of !10569. This series prepares wineandroid ioctl dispatch paths for execution on Java-originated threads that do not have a Wine TEB. It does three things: * pass JNIEnv explicitly through the ioctl dispatch path instead of relying on the global jni_env value in handlers * move native window registration from the APC/unixlib callback path to the direct JNI surface_changed callback on the Android UI thread * make logging in device.c context-aware so the same code remains safe both on Wine threads and on Java-only threads The JNIEnv change is a preparatory refactoring. It does not by itself complete the thread model transition, but it removes one of the major implicit dependencies on Wine thread context and makes follow-up changes smaller and easier to review. Moving native window registration to surface_changed removes another dependency on APC-based execution in the desktop process and puts window attachment on the thread that actually receives the Android surface callbacks. A shared lock is used to serialize registration against ioctl dispatch. Finally, logging in device.c is made context-aware. Upcoming changes will run parts of ioctl dispatch on Java-only threads, where the standard Wine TRACE/WARN/FIXME/ERR macros are unsafe because they depend on Wine TEB state. The local logging wrapper preserves normal Wine logging semantics on Wine threads and falls back to Android logging on non-Wine threads, allowing the existing code to keep using the standard logging macros without invasive per-call-site changes. Taken together, these patches remove several implicit Wine-thread-only assumptions from wineandroid ioctl dispatch code and prepare the path for running more of it directly on Java-side threads. -- v2: wineandroid: pass JNIEnv explicitly through ioctl handlers wineandroid: move native window registration to surface_changed wineandroid: make logging context-aware for upcoming Java-only ioctl dispatch https://gitlab.winehq.org/wine/wine/-/merge_requests/10710

2 4

[PATCH 0/1] MR10736: gdiplus: Don't add X margin on Center/Fat for GdipMeasureCharacterRanges
by Bartosz Kosiorek (＠gang65) April 27, 2026

April 27, 2026

We should shift X position only for StringAlignmentNear. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10736

4 9

[PATCH 0/3] MR10710: wineandroid: prepare ioctl dispatch for Java-only thread execution (part of !10569)
by Twaik Yont (＠twaik) April 27, 2026

April 27, 2026

Part of !10569. This series prepares wineandroid ioctl dispatch paths for execution on Java-originated threads that do not have a Wine TEB. It does three things: * pass JNIEnv explicitly through the ioctl dispatch path instead of relying on the global jni_env value in handlers * move native window registration from the APC/unixlib callback path to the direct JNI surface_changed callback on the Android UI thread * make logging in device.c context-aware so the same code remains safe both on Wine threads and on Java-only threads The JNIEnv change is a preparatory refactoring. It does not by itself complete the thread model transition, but it removes one of the major implicit dependencies on Wine thread context and makes follow-up changes smaller and easier to review. Moving native window registration to surface_changed removes another dependency on APC-based execution in the desktop process and puts window attachment on the thread that actually receives the Android surface callbacks. A shared lock is used to serialize registration against ioctl dispatch. Finally, logging in device.c is made context-aware. Upcoming changes will run parts of ioctl dispatch on Java-only threads, where the standard Wine TRACE/WARN/FIXME/ERR macros are unsafe because they depend on Wine TEB state. The local logging wrapper preserves normal Wine logging semantics on Wine threads and falls back to Android logging on non-Wine threads, allowing the existing code to keep using the standard logging macros without invasive per-call-site changes. Taken together, these patches remove several implicit Wine-thread-only assumptions from wineandroid ioctl dispatch code and prepare the path for running more of it directly on Java-side threads. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10710

3 11

[PATCH 0/1] MR10756: Uninstaller/Appwiz: update app list after installation
by Rose Hellsing (＠axtlos) April 27, 2026

April 27, 2026

Match the Uninstall workflow and make Installation synchronous to allow the app list to update immediately after installation finishes. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=48597 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10756

2 1

Re: [PATCH v4 0/1] MR10485: vbscript: Return proper errors for orphan Loop and Next keywords. - approved
by Jacek Caban (＠jacek) April 27, 2026

April 27, 2026

This merge request was approved by Jacek Caban. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10485

1 0

[PATCH v12 0/1] MR10314: vbscript: Convert string to number for comparison operators.
by Francis De Brabandere (＠francisdb) April 27, 2026

April 27, 2026

When one operand was a string and the other numeric, VBScript should have converted the string to a double before comparing. VarCmp did not do this automatically, resulting in incorrect string comparison. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=56281 -- v12: vbscript: Convert string to number for comparison operators. https://gitlab.winehq.org/wine/wine/-/merge_requests/10314

3 3