https://bugs.kde.org/show_bug.cgi?id=211031
--- Comment #15 from Austin English <austinengl...@gmail.com> --- As of wine-4.6, wine now supports building its dlls with mingw: https://source.winehq.org/git/wine.git/commitdiff/a3cf86a18446e9d7df4e045b83d3aa3d9193512a the plan is to eventually make that the default (not sure if there's a timeline for that yet, but I would guess that before next stable release is likely). When compiling wine with mingw support, stacktraces are pretty poor. As a comparison, here's the same test with native gcc (8.3.0) versus mingw (i686-w64-mingw32-gcc (Gentoo 9.2.0-r2 p3) 9.2.0). With dlls/comctl32/tests/edit.c in wine-4.20-213-gddec23013e native: ==10945== Invalid read of size 2 ==10945== at 0x4F0C401: LocalLock (memory.c:661) ==10945== by 0x63FAE1A: EDIT_LockBuffer (edit.c:1195) ==10945== by 0x640216D: EDIT_WindowProc (edit.c:4608) ==10945== by 0x55FB4A7: ??? (in /home/austin/wine-valgrind-no-mingw/dlls/user32/user32.dll.so) ==10945== by 0x55FBA94: call_window_proc (winproc.c:249) ==10945== by 0x55FCEFE: WINPROC_CallProcAtoW (winproc.c:609) ==10945== by 0x55FDB4D: WINPROC_call_window (winproc.c:956) ==10945== by 0x55C38CE: call_window_proc (message.c:2225) ==10945== by 0x55C6EB6: send_message (message.c:3294) ==10945== by 0x55C9B02: SendMessageA (message.c:3517) ==10945== by 0x5016FDF: test_EM_GETHANDLE (edit.c:3041) ==10945== by 0x5018F22: func_edit (edit.c:3411) ==10945== by 0x50A0581: run_test (test.h:637) ==10945== by 0x50A0E08: main (test.h:721) ==10945== Address 0x4952a58 is 16 bytes after a recently re-allocated block of size 32 alloc'd ==10945== at 0x7BC6656F: notify_alloc (heap.c:260) ==10945== by 0x7BC695B5: RtlAllocateHeap (heap.c:1725) ==10945== by 0x59177B7: create_family (freetype.c:1594) ==10945== by 0x59278A4: load_font_list_from_cache (freetype.c:1768) ==10945== by 0x59283B9: WineEngInit (freetype.c:4458) ==10945== by 0x5928CCD: DllMain (gdiobj.c:693) ==10945== by 0x594011C: __wine_spec_dll_entry (dll_entry.c:44) ==10945== by 0x7BC6AA95: ??? (in /home/austin/wine-valgrind-no-mingw/dlls/ntdll/ntdll.dll.so) ==10945== by 0x7BC6EC1A: MODULE_InitDLL (loader.c:1331) ==10945== by 0x7BC6EEBB: process_attach (loader.c:1425) ==10945== by 0x7BC6EDFD: process_attach (loader.c:1411) ==10945== by 0x7BC6EDFD: process_attach (loader.c:1411) ==10945== by 0x7BC6EDFD: process_attach (loader.c:1411) ==10945== by 0x7BC715B9: LdrInitializeThunk (loader.c:3782) ==10945== by 0x7BC9C556: attach_thread (signal_i386.c:2746) ==10945== by 0x7BC991C3: ??? (in /home/austin/wine-valgrind-no-mingw/dlls/ntdll/ntdll.dll.so) mingw: ==15488== Invalid read of size 2 ==15488== at 0x7BC74A86: RtlInitNlsTables (locale.c:760) ==15488== by 0x7125EB11: ??? ==15488== by 0x7126254F: ??? ==15488== by 0x71262871: ??? ==15488== by 0x7BC6AAD5: ??? (in /home/austin/wine-valgrind-mingw/dlls/ntdll/ntdll.dll.so) ==15488== by 0x7BC6EC5A: MODULE_InitDLL (loader.c:1331) ==15488== by 0x7BC6EEFB: process_attach (loader.c:1425) ==15488== by 0x7BC6EE3D: process_attach (loader.c:1411) ==15488== by 0x7BC6EE3D: process_attach (loader.c:1411) ==15488== by 0x7BC715F9: LdrInitializeThunk (loader.c:3782) ==15488== by 0x7BC9D12A: attach_thread (signal_i386.c:2746) ==15488== by 0x7BC99D97: ??? (in /home/austin/wine-valgrind-mingw/dlls/ntdll/ntdll.dll.so) ==15488== Address 0x2 is not stack'd, malloc'd or (recently) free'd ==15488== interestingly, mingw shows RtlInitNlsTables while gcc doesn't (but that's likely it, since RtlInitNlsTables is in dlls/ntdll/locale.c at line 760 and that's the ??? in the gcc output. -- You are receiving this mail because: You are watching all bug changes.
