To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=44627





------- Additional comments from [EMAIL PROTECTED] Fri Mar 11 02:18:53 -0800 
2005 -------
Hi Michael, Hennes,

Okay, I now seem to understand what you're at: consistency between the w32 and
unx implementations.

And, yes, I agree that they are currently inconstent, as the w32 implementation
as it is now corresponds to the plain pthread_cond_broadcast() of a
pthread_cond_t (plain signaling of an event object), but does not care about any
postcondition (like the unx implementation does; we could insert an "assert
(pCond->m_State)" at the end of osl_waitCondition(), just before releasing the
associated mutex).

Thus, assuming that an oslCondition should model the Win32 behaviour of a simple
event object, the unx implementation must indeed be changed as Michael proposed
(i.e. replace "while ( ! pCond->m_State )" with "if ( ! pCond->m_State)"). That
would of course render an oslCondition unusable for the producer / consumer
scenario that I mentioned recently (an additional check for the application
"state" is then necessary).

Just fixing the inconsistency could of course also be achieved through
introducing a "pCond->m_State" into the w32 implementation, and handle that as
in the unx case.

In order to make both, w32 and unx, implementations conform to the documentation
("unblock all threads"), Michaels proposal seems to be the correct one. So,
please go ahead.

Hope that helps,
Matthias


---------------------------------------------------------------------
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