To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=93051





------- Additional comments from [EMAIL PROTECTED] Wed Aug 27 12:03:10 +0000 
2008 -------
> jb wrote:
> When I left configmgr it was very much threading-capable. The UnoApiLock was
> obviously introduced by mmeeks in the configrefactor01 CWS.

nonsense - it was riddled with lock takes & releases, certainly - but any
threaded code actively using the configmgr deadlocked or crashed within a few
seconds: cf. my regression test. This is now no longer the case - and yes, the
threading model is -far- simpler now.

Of course - that means it can sometimes deadlock - but at least we get a clear
view of the interactions between pieces of code; and the existing lock hierarchy
problems around the place.

The UnoApiLock is a simple code lock; we count take / releases on it - and we
drop this lock as we emit callbacks & call out of the code [ at least in theory
of course ].

All that should be necessary is to create an UnoApiLockReleaser over the service
factory instantiation; and isolate the interaction carefully to make sure it's 
safe.


---------------------------------------------------------------------
Please do not reply to this automatically generated notification from
Issue Tracker. Please log onto the website and enter your comments.
http://qa.openoffice.org/issue_handling/project_issues.html#notification

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to