- Wine-GitLab - list.winehq.org

[PATCH v12 0/3] MR8030: wineserver: Allow NtOpenFile with read-only files.
by Joel Holdsworth (＠jhol) May 19, 2026

May 19, 2026

Co-authored-by: Bernhard Übelacker <bernhardu(a)mailbox.org> -- v12: server: Don't require Unix read or write permission for reading or writing only attributes. https://gitlab.winehq.org/wine/wine/-/merge_requests/8030

3 3

[PATCH v4 0/2] MR10889: ole32: Validate cached clipboard window handle in get_clipbrd_window.
by Aaron Yourk (＠ayourk) May 19, 2026

May 19, 2026

The OLE clipboard caches a per-process window handle. When the STA thread that created this window terminates, the window is destroyed but the stale handle remains cached. Subsequent OleSetClipboard calls use this stale handle, causing OpenClipboard to fail with CLIPBRD_E_CANT_OPEN. Check whether the cached window is still valid using IsWindow() before returning it. If the window has been destroyed, clear the cache so a fresh window is created on the current thread. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59519 -- v4: ole32: Validate cached clipboard window handle in get_clipbrd_window. https://gitlab.winehq.org/wine/wine/-/merge_requests/10889

3 4

[PATCH 0/3] MR10951: wineandroid: Fix focus, activation and duplicate input issues
by Twaik Yont (＠twaik) May 19, 2026

May 19, 2026

These changes fix several wineandroid focus and input handling bugs I discovered while tinkering with Wine Android, particularly while testing VirtualBox on Android 8.1. Changes included: - keep keyboard focus on whole Wine views instead of client rendering surfaces - request focus for the desktop view immediately after startup - ignore duplicate primary button motion events reported by some Android devices Together these changes fix issues such as: - keyboard input not working until focus changes externally - duplicate mouse button events breaking capture state - inconsistent focus behavior between client and whole views -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10951

2 3

[PATCH v10 0/1] MR10660: uniscribe: Correct possible buffer over/under runs
by Aric Stewart (＠aricstewart) May 19, 2026

May 19, 2026

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59532 -- v10: uniscribe: Correct possible buffer over/under runs https://gitlab.winehq.org/wine/wine/-/merge_requests/10660

2 3

Re: [PATCH v9 0/1] MR10660: uniscribe: Correct possible buffer over/under runs
by Aric Stewart (＠aricstewart) May 19, 2026

May 19, 2026

On Tue May 19 08:09:55 2026 +0000, Huw Davies wrote: > Do we need to worry about `next_glyph_index < 0`? > Also, opening brace on new line seems to be the majority style here. No, next_glyph_index is an unsigned int, so if it would be less than 0 it would instead be an extremely large number that will definitely be > glyph_count -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10660#note_140608

1 0

[PATCH v5 0/1] MR10937: vbscript: Parse double literals via ucrtbase _wcstod_l.
by Francis De Brabandere (＠francisdb) May 19, 2026

May 19, 2026

The custom pow()-based real-number conversion was historical: when this code was written, wine could not link against ucrtbase, so the lexer reimplemented the conversion. That constraint is gone, and the custom code had two correctness gaps - subnormal literals (e.g. 1e-309) rounded to zero because pow(10, 309) overflows to infinity, and values near DBL_MAX could round through the same infinity edge. Hand the literal span to _wcstod_l with a C locale instead, which handles rounding, subnormals, and overflow per IEEE 754 - and is locale-invariant by construction, matching VBScript source semantics where the decimal separator is always '.' regardless of runtime SetLocale. Above-DBL_MAX literals (1e309, 1.8e308) continue to raise err 1031 as before. alternative for !10898 supersedes !7591 -- v5: vbscript: Parse double literals via ucrtbase _wcstod_l. https://gitlab.winehq.org/wine/wine/-/merge_requests/10937

2 3

Re: [PATCH v4 0/1] MR10937: vbscript: Parse double literals via ucrtbase _wcstod_l.
by Francis De Brabandere (＠francisdb) May 19, 2026

May 19, 2026

On Tue May 19 14:55:35 2026 +0000, Francis De Brabandere wrote: > changed this line in [version 4 of the diff](/wine/wine/-/merge_requests/10937/diffs?diff_id=268696&start_sha=7b6fe67b96ef8f1c9f10cca2f68ecd8fceff9989#e19d7f25e61cb09908e1164993ab8330b56e13e3_353_354) Native check: cscript on the Win10 VM accepts literals up to at least 5000 chars -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10937#note_140595

1 0

[PATCH v6 0/1] MR10900: vbscript: Parse octal numeric literals (&O...).
by Francis De Brabandere (＠francisdb) May 19, 2026

May 19, 2026

Native cscript accepts &O0, &O17, &O177777, etc. as octal integer literals. Wine was emitting err 1002 (syntax error) for any &O prefix because the lexer only recognised &H. Add an oct_to_int helper and parse_oct_literal mirroring parse_hex_literal, allowing up to 32-bit values; overflow checks use an ULONGLONG accumulator since 11-digit octal can exceed 32 bits. -- v6: vbscript: Parse octal numeric literals (&O...). https://gitlab.winehq.org/wine/wine/-/merge_requests/10900

3 5

[PATCH v3 0/1] MR10937: vbscript: Parse double literals via ucrtbase _wcstod_l.
by Francis De Brabandere (＠francisdb) May 19, 2026

May 19, 2026

The custom pow()-based real-number conversion was historical: when this code was written, wine could not link against ucrtbase, so the lexer reimplemented the conversion. That constraint is gone, and the custom code had two correctness gaps - subnormal literals (e.g. 1e-309) rounded to zero because pow(10, 309) overflows to infinity, and values near DBL_MAX could round through the same infinity edge. Hand the literal span to _wcstod_l with a C locale instead, which handles rounding, subnormals, and overflow per IEEE 754 - and is locale-invariant by construction, matching VBScript source semantics where the decimal separator is always '.' regardless of runtime SetLocale. Above-DBL_MAX literals (1e309, 1.8e308) continue to raise err 1031 as before. alternative for !10898 supersedes !7591 -- v3: vbscript: Parse double literals via ucrtbase _wcstod_l. https://gitlab.winehq.org/wine/wine/-/merge_requests/10937

3 2

[PATCH v10 0/2] MR10895: vbscript: Resolve UBound/LBound default member and match native errors
by Francis De Brabandere (＠francisdb) May 19, 2026

May 19, 2026

Resolve VT_DISPATCH arguments via DISPID_VALUE so an object whose default member returns an array works, and map the failure modes to native error codes (91 for Nothing, 438 for no default, 450 for default needing args, 13 for default returning a scalar). Scalar arguments collapse to err 13. -- v10: vbscript: Match native error codes for non-array arguments to UBound/LBound. https://gitlab.winehq.org/wine/wine/-/merge_requests/10895

3 4