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

[PATCH v3 0/2] MR10271: Fixed thread safety issues with NtUserRegistClassExWOW and NtUserUnregisterClass.
by Yongjie Yao (＠yaoyongjie) March 24, 2026

March 24, 2026

Kingsoft Typing Pass failed to start probabilistically due to a deadlock between two threads registering and unregistering the same class. The replication scenario is as follows: |A thread (registration class) | B thread (deregistration class)| |-------------- |---------------| | NtUserRegisterClassExWOW | NtUserUnregisterClass | | create_class | -- | | -- | destroy_class | | find_shared_session_object | -- | | find failed, destroy_class | -- | | free class | -- | | -- | user_lock | | -- | remove class from list | | -- | throw exception(class has freeed by thread A) | Thread B threw an exception and will not call user_unlock. The upper layer application handled the memory access exception thrown by thread B, but due to If user_unlock is not called, the lock will remain occupied, and other threads will continue to wait when calling user_rock. -- v3: Win32u: Fixed thread safety issues with NtUserRegistClassExWOW and NtUserUnregisterClass. user32/tests: Test the thread safety of NtUserRegisterClassExWOW and NtUserUnregisterClass. https://gitlab.winehq.org/wine/wine/-/merge_requests/10271

3 3

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Nikolay Sivov (＠nsivov) March 24, 2026

March 24, 2026

On Tue Mar 24 09:56:11 2026 +0000, Hans Leidekker wrote: > Should the last parameter to FCICompressorCreate() be HFCICOMPRESSOR * ? Yes, but it's called PCOMPRESSOR_HANDLE. Types should be revised too of course, not only names. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133519

1 0

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Hans Leidekker (＠hans) March 24, 2026

March 24, 2026

On Tue Mar 24 09:29:41 2026 +0000, Nikolay Sivov wrote: > Yes, existing exports should of course stay as is. Newer exports are > different, and we can't change that. Should the last parameter to FCICompressorCreate() be HFCICOMPRESSOR * ? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133518

1 0

[PATCH 0/8] MR10435: msxml3/tests: Add ISupportErrorInfo tests for DOCTYPE nodes.
by Nikolay Sivov (＠nsivov) March 24, 2026

March 24, 2026

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10435

2 8

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Hans Leidekker (＠hans) March 24, 2026

March 24, 2026

On Tue Mar 24 09:13:59 2026 +0000, Nikolay Sivov wrote: > Those should use stdcall. Parameters should be specified for implemented functions. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133517

1 0

[PATCH 0/1] MR10434: ntdll: NtFlushVirtualMemory correct parameter
by Alistair Leslie-Hughes (＠alesliehughes) March 24, 2026

March 24, 2026

msdn: ZwFlushVirtualMemory states the last parameter is PIO_STATUS_BLOCK. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10434

2 1

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Nikolay Sivov (＠nsivov) March 24, 2026

March 24, 2026

On Tue Mar 24 09:23:40 2026 +0000, Chris Kadar wrote: > I used > https://learn.microsoft.com/en-us/windows/win32/api/compressapi/nf-compress… > for function definitions, but for function names I tried to match the > existing style. > Looking at it now, the existing style was defined by the fci.h header, > so that doesn't really make sense. I wanted to keep the change minimal > and avoid adding files, but perhaps I should do that here after all? Yes, existing exports should of course stay as is. Newer exports are different, and we can't change that. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133514

1 0

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Chris Kadar (＠nspitko) March 24, 2026

March 24, 2026

On Tue Mar 24 09:13:59 2026 +0000, Nikolay Sivov wrote: > If it's not important to get application running, it's better to keep > stub to a minimum. For instance here, we don't know if this is correct, > without tests. I believe this is the minimal stub for this game to work. The call pattern I observed is that it will call Compress first with a 0 byte input buffer, expecting to get ERROR_INSUFFICIENT_BUFFER back and reads the output data size to produce a new buffer. It then re-calls the function with the corrected size and expected data written to it. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133513

1 0

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Chris Kadar (＠nspitko) March 24, 2026

March 24, 2026

On Tue Mar 24 09:13:59 2026 +0000, Nikolay Sivov wrote: > Is it defined like that in some older SDK? In current SDK there is no > such symbol, and compress/decompress functions are provided with compressapi.h. I used https://learn.microsoft.com/en-us/windows/win32/api/compressapi/nf-compress… for function definitions, but for function names I tried to match the existing style. Looking at it now, the existing style was defined by the fci.h header, so that doesn't really make sense. I wanted to keep the change minimal and avoid adding files, but perhaps I should do that here after all? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133512

1 0

Re: [PATCH v2 0/1] MR10352: cabinet: Add stubs for compressor/decompressor
by Nikolay Sivov (＠nsivov) March 24, 2026

March 24, 2026

Nikolay Sivov (@nsivov) commented about dlls/cabinet/fci.c: > + PVOID buff, > + SIZE_T buffSize, > + PSIZE_T compDataSize) > +{ > + FIXME("stub\n"); > + > + *compDataSize = dataSize; > + > + if( buffSize < dataSize ) > + { > + SetLastError( ERROR_INSUFFICIENT_BUFFER ); > + return FALSE; > + } > + memcpy(buff, data, dataSize); > + return TRUE; > +} If it's not important to get application running, it's better to keep stub to a minimum. For instance here, we don't know if this is correct, without tests. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10352#note_133509

1 0