[Bug 55333] New: ws2_32:sock - test_connect_events() sometimes fails in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55333

Bug ID: 55333
           Summary: ws2_32:sock - test_connect_events() sometimes fails in
                    Wine
           Product: Wine
           Version: 7.21
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Keywords: source, testcase
          Severity: normal
          Priority: P2
         Component: winsock
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

It seems to get FD_WRITE *after* FD_CONNECT, resulting in the following failure
messages:

sock.c:6408: Test failed: got events 0x10
sock.c:6409: Test failed: expected timeout
sock.c:6409: Test failed: got events 0x2

I don't immediately know how this can happen. I'm pretty sure we depend on
POLLOUT to trigger both.

[Bug 55333] ws2_32:sock - test_connect_events() sometimes fails in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55333

François Gouget  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
    Regression SHA1|                            |01afb21bf2b3a9c6b03847b03a1
                   |                            |ac4a572ed49a3
                 CC|                            |[e-mail]
           Keywords|                            |regression

--- Comment #1 from François Gouget  ---
A bisect shows that this started with the commit below:

commit 01afb21bf2b3a9c6b03847b03a1ac4a572ed49a3
Author:     Zebediah Figura 
AuthorDate: Sat Jul 22 01:44:06 2023 -0500

ws2_32: Move the event reset in WSAEnumNetworkEvents() to the server.

So that it is done atomically with retrieving events.

Wine-Bug: https://bugs.winehq.org//show_bug.cgi?id=52474

[Bug 55333] ws2_32:sock - test_connect_events() sometimes fails in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55333

--- Comment #2 from Zeb Figura  ---
(In reply to Zeb Figura from comment #0)
> I don't immediately know how this can happen. I'm pretty sure we depend on
> POLLOUT to trigger both.

Ugh, rather simply, as it happens. After a7577014faed we don't report the
events at the same time anymore.

The regression commit is probably not *really* the source of the regression. I
haven't thought through it fully, but it probably just tweaks timing a bit so
that it's easier to run into this bug.

It should be easy to fix, though.

[Bug 55333] ws2_32:sock - test_connect_events() sometimes fails in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55333

Zeb Figura  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
      Fixed by SHA1|                            |dbe7788817a6e0192ce953e6b9a
                   |                            |32d8e9318d2df
             Status|NEW                         |RESOLVED

--- Comment #3 from Zeb Figura  ---
Assuming fixed by
https://source.winehq.org/git/wine.git/commitdiff/dbe7788817a6e0192ce953e6b9a32d8e9318d2df;
let's reopen if this shows up again.

[Bug 55333] ws2_32:sock - test_connect_events() sometimes fails in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55333

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #4 from Alexandre Julliard  ---
Closing bugs fixed in 8.15.