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

[PATCH 0/1] MR10789: msvcp120: Ensure that segments are not being allocated when segments table is enlarged.
by Piotr Caban (＠piotr) April 30, 2026

April 30, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10789

2 2

[PATCH v3 0/1] MR10782: win32u: Stub NtUserInitializeTouchInjection.
by Rémi Bernon (＠rbernon) April 30, 2026

April 30, 2026

Supersedes https://gitlab.winehq.org/wine/wine/-/merge_requests/10778 -- v3: win32u: Stub NtUserInitializeTouchInjection. https://gitlab.winehq.org/wine/wine/-/merge_requests/10782

3 2

[PATCH 0/1] MR10787: ntdll: Use ESR in the ARM64 signal handler when possible.
by Jacek Caban (＠jacek) April 30, 2026

April 30, 2026

Fixes SIGTRAP handling on macOS, which does not set si_code but provides a valid ESR. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10787

2 2

[PATCH v2 0/2] MR10785: msvcrt: Only reject negative timestamps with EINVAL in ctime_s
by Hendrik Borchardt (＠hborchardt) April 30, 2026

April 30, 2026

0 needs to be an allowed value, and negative values should not halt the program but just return EINVAL. This behavior has been observed with the following python script: ```python import ctypes ucrt = ctypes.CDLL("ucrtbase.dll") buf = ctypes.create_string_buffer(26) for i in [1, 0, -1]: print(f"_ctime64_s for time {i}:") t = ctypes.c_int64(i) err = ucrt._ctime64_s(buf, ctypes.c_size_t(26), ctypes.byref(t)) if err == 0: print(buf.value.decode()) else: print(f"Error: {err}") for i in [1, 0, -1]: print(f"_ctime32_s for time {i}:") t = ctypes.c_int32(i) err = ucrt._ctime32_s(buf, ctypes.c_size_t(26), ctypes.byref(t)) if err == 0: print(buf.value.decode()) else: print(f"Error: {err}") # _ctime64_s for time 1: # Thu Jan 1 01:00:01 1970 # # _ctime64_s for time 0: # Thu Jan 1 01:00:00 1970 # # _ctime64_s for time -1: # Error: 22 # # _ctime32_s for time 1: # Thu Jan 1 01:00:01 1970 # # _ctime32_s for time 0: # Thu Jan 1 01:00:00 1970 # # _ctime32_s for time -1: # Error: 22 ``` The bug https://bugs.winehq.org/show_bug.cgi?id=57261 is solved by this, because mongodb tries to pretty-print a 0 timestamp, and sets up its own invalid_parameter_handler, which crashes due to NULL arguments. -- v2: msvcrt: Only reject negative timestamps with EINVAL in ctime_s msvcrt/tests: Test that ctime_s succeeds for timestamp 0 https://gitlab.winehq.org/wine/wine/-/merge_requests/10785

3 5

[PATCH 0/5] MR10786: msxml3: Pass parsed DTD to the DOM.
by Nikolay Sivov (＠nsivov) April 30, 2026

April 30, 2026

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10786

2 5

[PATCH v12 0/6] MR10775: vbscript: Fix and expand test coverage for BSTR comparison. (alternate)
by Francis De Brabandere (＠francisdb) April 30, 2026

April 30, 2026

Alternate version of !10766 This replaces runtime VT_RESERVED tracking with compile-time is_literal_expr at the AST level: comparison opcodes (equal/nequal/gt/gteq/lt/lteq/case) carry a flag arg with bit 0 = left-literal, bit 1 = right-literal, and var_cmp dispatches on those flags. No runtime stripping, no OP_strip_reserved, and Const references stay non-literal because they're EXPR_MEMBER nodes even though the value is inlined. -- v12: vbscript: Reject Automation types not supported in VBScript on compare. vbscript/tests: Cover BSTR comparison with dispatch-only numeric VTs. https://gitlab.winehq.org/wine/wine/-/merge_requests/10775

2 6

[PATCH 0/2] MR10785: msvcrt: Only reject negative timestamps with EINVAL in ctime_s
by Hendrik Borchardt (＠hborchardt) April 30, 2026

April 30, 2026

0 needs to be an allowed value, and negative values should not halt the program but just return EINVAL. This behavior has been observed with the following python script: ```python import ctypes ucrt = ctypes.CDLL("ucrtbase.dll") buf = ctypes.create_string_buffer(26) for i in [1, 0, -1]: print(f"_ctime64_s for time {i}:") t = ctypes.c_int64(i) err = ucrt._ctime64_s(buf, ctypes.c_size_t(26), ctypes.byref(t)) if err == 0: print(buf.value.decode()) else: print(f"Error: {err}") for i in [1, 0, -1]: print(f"_ctime32_s for time {i}:") t = ctypes.c_int32(i) err = ucrt._ctime32_s(buf, ctypes.c_size_t(26), ctypes.byref(t)) if err == 0: print(buf.value.decode()) else: print(f"Error: {err}") # _ctime64_s for time 1: # Thu Jan 1 01:00:01 1970 # # _ctime64_s for time 0: # Thu Jan 1 01:00:00 1970 # # _ctime64_s for time -1: # Error: 22 # # _ctime32_s for time 1: # Thu Jan 1 01:00:01 1970 # # _ctime32_s for time 0: # Thu Jan 1 01:00:00 1970 # # _ctime32_s for time -1: # Error: 22 ``` The bug https://bugs.winehq.org/show_bug.cgi?id=57261 is solved by this, because mongodb tries to pretty-print a 0 timestamp, and sets up its own invalid_parameter_handler, which crashes due to NULL arguments. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10785

2 2

[PATCH v11 0/6] MR10775: vbscript: Fix and expand test coverage for BSTR comparison. (alternate)
by Francis De Brabandere (＠francisdb) April 30, 2026

April 30, 2026

Alternate version of !10766 This replaces runtime VT_RESERVED tracking with compile-time is_literal_expr at the AST level: comparison opcodes (equal/nequal/gt/gteq/lt/lteq/case) carry a flag arg with bit 0 = left-literal, bit 1 = right-literal, and var_cmp dispatches on those flags. No runtime stripping, no OP_strip_reserved, and Const references stay non-literal because they're EXPR_MEMBER nodes even though the value is inlined. -- v11: vbscript: Reject Automation types not supported in VBScript on compare. https://gitlab.winehq.org/wine/wine/-/merge_requests/10775

2 7

[PATCH v55 0/2] MR10295: Fix ImageGetDigestStream
by Trent Waddington (＠trent.waddington) April 30, 2026

April 30, 2026

ImageGetDigestStream is used to hash the "important" parts of PE files. After rigorous experimentation we discovered these changes were sufficient to get identical hashes (and thereby signatures) on Windows and Wine. -- v55: imagehlp: Fix ImageGetDigestStream https://gitlab.winehq.org/wine/wine/-/merge_requests/10295

3 4

[PATCH v10 0/5] MR10775: Draft: vbscript: Fix and expand test coverage for BSTR comparison. (alternate)
by Francis De Brabandere (＠francisdb) April 30, 2026

April 30, 2026

Alternate version of !10766 This replaces runtime VT_RESERVED tracking with compile-time is_literal_expr at the AST level: comparison opcodes (equal/nequal/gt/gteq/lt/lteq/case) carry a flag arg with bit 0 = left-literal, bit 1 = right-literal, and var_cmp dispatches on those flags. No runtime stripping, no OP_strip_reserved, and Const references stay non-literal because they're EXPR_MEMBER nodes even though the value is inlined. -- v10: vbscript/tests: Cover BSTR comparison with dispatch-only numeric VTs. https://gitlab.winehq.org/wine/wine/-/merge_requests/10775

2 5