[Bug 15462] Wine blocking on exit when stdout and stderr are piped

[wine-bugs at winehq.org]

http://bugs.winehq.org/show_bug.cgi?id=15462

Andrey Turkin <andrey.turkin at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andrey.turkin at gmail.com

--- Comment #6 from Andrey Turkin <andrey.turkin at gmail.com> 2013-05-21 02:29:00 CDT ---
Here's that happens:
1) One wine process forks off wineserver which would leave stderr open in case
it needs to dump some traces or error messages; this stderr handle (inherited
from wine process) is connected to write side of a pipe which tee reads
2) wine process for another script instance uses this wineserver for its
operation
3) wine process that initially started wineserver exits; wineserver still
lingers for few more seconds; tee waits for all the writers to go away
4) second wine process in another script instance starts up, connects to the
wineserver and runs to completion
5) wineserver waits for few seconds again, then exits
6) tee exits, second wine process is started (starting new wineserver)

Basically the question here is when wineserver should close its inherited
stderr when implicitly started in background by wine (right away from the
start, or when program that spawned wineserver exits, or when all non-system
programs started by that original program exited, or never, or something else). 
Also there is an obvious workaround described by Chris - just start some wine
application beforehand to make sure wineserver is there.

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.