[Pharo-project] Issue 1053: Transcript updates from foreground process not displayed in Transcript window
Ready for review: http://code.google.com/p/pharo/issues/detail?id=1053
thanks On Aug 10, 2009, at 3:14 AM, Dale Henrichs wrote:
Ready for review:
http://code.google.com/p/pharo/issues/detail?id=1053
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Hi, if you fork processes at a higher priority, they still throw the error in issue 1053. In core 10428 I more often get Instances of UndefinedObject are not indexable. It is trying to send size to a MultiNewParagraph while it is still nil. By the time the debugger opens, the object is there. Test case attached, although I can't get it to catch the error. Use with ***caution***. In the one click image it often freezes the image. http://n2.nabble.com/file/n3565757/NSThreadsafeTranscriptPriorityTest.st NSThreadsafeTranscriptPriorityTest.st It seems like this may be the case Dale identified- "I assume that with a sufficiently complex test case, involving heavy background window updates while using some of the operations that call #displayWorldSafely a problem would be revealed, but there's no guarantee that a mutex would be the solution". Cheers, ...Stan Dale Henrichs wrote:
Ready for review:
http://code.google.com/p/pharo/issues/detail?id=1053
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
-- View this message in context: http://n2.nabble.com/Issue-1053-Transcript-updates-from-foreground-process-n... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.
Thanks may be I should not have tried to make the Transcript threadsafe. And may be one of these days it would be great to get a cleaner model for morphic threading. stef
Hi, if you fork processes at a higher priority, they still throw the error in issue 1053. In core 10428 I more often get Instances of UndefinedObject are not indexable. It is trying to send size to a MultiNewParagraph while it is still nil. By the time the debugger opens, the object is there.
Test case attached, although I can't get it to catch the error. Use with ***caution***. In the one click image it often freezes the image.
http://n2.nabble.com/file/n3565757/NSThreadsafeTranscriptPriorityTest.st NSThreadsafeTranscriptPriorityTest.st
It seems like this may be the case Dale identified- "I assume that with a sufficiently complex test case, involving heavy background window updates while using some of the operations that call #displayWorldSafely a problem would be revealed, but there's no guarantee that a mutex would be the solution".
Cheers, ...Stan
Dale Henrichs wrote:
Ready for review:
http://code.google.com/p/pharo/issues/detail?id=1053
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
-- View this message in context: http://n2.nabble.com/Issue-1053-Transcript-updates-from-foreground-process-n... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Wed, Sep 2, 2009 at 11:43 AM, Stan Shepherd <stan.shepherd414@gmail.com>wrote:
Test case attached, although I can't get it to catch the error. Use with ***caution***. In the one click image it often freezes the image.
I note for these cases where catching the error with an debug dialog seems impossible, goodness knows we saw lots of that in Sophie via Tweak, then try something like Smalltalk logError:'Fatal error' inContext: thisContext to: 'SqueakDebug.log'.]] This then logs the context to a log file, for later review. For evil UI related bugs this can be of assistance. -- =========================================================================== John M. McIntosh <johnmci@smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================
participants (4)
-
Dale Henrichs -
John McIntosh -
Stan Shepherd -
Stéphane Ducasse