[Bug 54061] New: shell32:shelllink - test_load_save() crashes randomly in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54061

Bug ID: 54061
           Summary: shell32:shelllink - test_load_save() crashes randomly
                    in Wine
           Product: Wine
           Version: unspecified
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: shell32
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

shell32:shelllink - test_load_save() crashes randomly in Wine:

shelllink.c:847: Test failed: SHCreateStreamOnFileW failed 80070020
shelllink.c:849: Test failed: IPersistFile_Save failed (0x00000000)
shelllink.c:851: Test failed: IPersistFile_Load failed (0x00000000)
Unhandled exception: page fault on read access to 0x00000000 in 32-bit code
(0x0042321a).
[...useless backtrace...]

See https://test.winehq.org/data/patterns.html#shell32:shelllink

Where:
0x80070020 == HRESULT_FROM_WIN32(ERROR_SHARING_VIOLATION)

So this may still have a similar cause as bug 28290, i.e. an invalid sharing
mode somewhere.

[Bug 54061] shell32:shelllink - test_load_save() crashes randomly in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54061

François Gouget  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |source, testcase

[Bug 54061] shell32:shelllink - test_load_save() crashes randomly in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54061

--- Comment #1 from François Gouget  ---
The failing tests were introduced on 2022-09-10 by the commit below:

commit f49a30167a945062e197dcbd8848c9f46ef3da26
Author:     Alexandre Julliard 
AuthorDate: Thu Sep 1 15:59:41 2022 +0200

shell32: Fix sharing modes when loading/saving link files.

The first failure happened on 2022-09-30 and it has happened about 3 times per
month since. Despite that it's somewhat hard to reproduce as it seems to depend
on the CPU / IO load. So it's plausible that this was a preexisting issue or
that it was introduced by this commit but it's also not very clear :-(