Re: Issue 5092 in pharo: Debugger in morphic thread doesn't show up

Comment #1 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

simple cs to reproduce it.

Load it and evaluate "TestMorph new openInWorld"

Attachments:
        test.st  255 bytes


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #2 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

BTW, if you put a halt, lots of debugger windows open one after another


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #3 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

hmm, it seems that Morph>>fullDrawOn: is eating the exceptions :)

It does:

[
stuff
stuff
] on: Error do: [:err |
     self setProperty: #errorOnDraw toValue: true.
     ^self drawErrorOn: aCanvas.
]

I tried replacing that by

[
stuff
stuff
] on: Error do: [:err |
     self setProperty: #errorOnDraw toValue: true.
     self drawErrorOn: aCanvas.
     err pass
]

And it brings the debugger :).

a CS soon :P


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #4 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

Ok, I figured out that catching the exception like:

on: Error, Halt do: ...

behaves correctly on a debugger, but I do not fully understand why yet :).

Cs with fix proposed

Attachments:
        Issue 5092: Debugger in morphic thread doesn't show up.1.cs  1.1 KB


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: FixReviewNeeded

Comment #5 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

(No comment was entered for this change.)


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #6 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

Hmm, looking at older vm's, it does not look a good solution, but I do not  
understand why the default error handler of the current process is called  
after all.  Or where the debugger is raised :(.


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: Accepted

Comment #7 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

(No comment was entered for this change.)


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #8 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

*older vms -> older images


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #9 on issue 5092 by [hidden email]: Debugger in morphic thread  
doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

Recursion.

If when an error is made during morph drawing, you would launch a debugger  
on the UI thread, that would start a new UI thread to display a debugger on  
the suspended one, which would* make an error during morph drawing, leading  
to launching  a debugger, starting a new UI thread...

* would here meaning limited cases, as when the morph with drawing errors  
is used for the debugger, or part of opened debugger background, which  
needs to be invalidated. Important thing is, passing the error to get a  
Debugger opened can't guarantee you won't end up in a pickle.




_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Labels: -Milestone-1.4

Comment #10 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

this problem is there forever (since squeak 1.X when morhic was introduced)

Thus: not a show stopper to hold up release of 1.4


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #11 on issue 5092 by [hidden email]: Debugger in morphic  
thread doesn't show up
http://code.google.com/p/pharo/issues/detail?id=5092

wouldnt it be possible to call the debugger directly in the errorhandler of  
fullDrawOn: ?
eg something like this:

on: (Error,Halt) do: [:err|
        self setProperty: #errorOnDraw toValue: true.
        (err isKindOf: Halt)ifTrue: [err messageText:'Halt' ].
        Processor activeProcess debug: err signalerContext title: err messageText.
        ]

probably i dont really understand the recursion argument: of course it  
makes sense if the morph is part of the debugger. but if the morph gets  
invalidated as part of the background #errorOnDraw is already set and the  
drawing moves out of most problematic areas early on (eg it wont reach  
drawOn:, or this errorhandler), or?
werner


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker