Issue 4200 in pharo: Build of Pharo Full does not terminate

Status: Accepted
Owner: [hidden email]
CC: marianopeck
Labels: Milestone-1.3-DevImage

New issue 4200 by [hidden email]: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

Build of Pharo Full does not terminate

https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.3/

already 2 hours running. Never managed to build starting 11.05.2011


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

Updates:
        Status: Closed

Comment #1 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

.... took 2 hr 41 min


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

Updates:
        Cc: [hidden email]

Comment #2 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

weird...I think it is the Compiler recompileAll that suddently it takes  
ages.


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

Comment #3 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

Try this and see if it will work faster.
:)

The problem is that to be fast, means losing source code.
But i think its okay, since we're speaking about methods which are  
deinstalled from classes.

Attachments:
        erasing-trailers.1.cs  2.1 KB


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

Comment #4 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

oh crap.. why ChangeSet can't preserve the orders in which i made the  
changes?

i split it on phases, because if you file-in the previous changeset it will  
shot into own feet. :)

Attachments:
        erasing-trailers-phase1.1.cs  1.1 KB
        erasing-trailers-phase2.1.cs  1.1 KB


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

Updates:
        Status: FixToInclude

Comment #5 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

You rock Igor :)
Now it is much faster. Almost as it was before I could say. What I don't  
understand how can the cange of the other day impact so much. Because this  
was working fast just before that...

Marcus, I put it ready to integrate :)


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

Updates:
        Labels: -Milestone-1.3-DevImage Milestone-1.3

Comment #6 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

(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 #7 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

Issue 4201 has been merged into this issue.


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

Comment #8 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

mmmm  no wait don't include


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

Updates:
        Status: Accepted

Comment #9 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

After doing these file ins, loading stuff from Monticello let me some CMs  
without source!!  For example, I load Fuel Package, and I have this method  
look like this:

FLCompiledMethodClusterSerializer >> referencesOf: t1 do: t2
        1
                to: t1 numLiterals
                do: [:t3 | t2
                                value: (t1 literalAt: t3)]


Instead of

referencesOf: anObject do: aBlock

        1 to: anObject numLiterals
                do: [ :index | aBlock value: (anObject literalAt: index) ]


So....it is being decompiled...it lost its sourceCode! :(


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

Comment #10 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

is monticello uninstalls methods and then installing them back?
because if it is, then it will lose the source code.

Please test what happens by removing sends #dropSourcePointer one by one.


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

Comment #11 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

I've checked. Monticello does a dropSourcePointer, but ONLY (as I can see)  
when using Trait with methods that overwrite methods.

Example, in Fuel, FLObjectClusterSerializer uses a trait called  
FLTClassParameter which implements #=
See attached stacktrace.

Attachments:
        DropTrait.rtf  32.8 KB


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

Comment #12 on issue 4200 by [hidden email]: Build of Pharo Full does  
not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

sorry, i can't really see what happens by reading this trace.
So, if Trait replacing the existing method in class, why that old method  
should preserve its source code?

The comment of #addTraitSelector:withMethod:
explicitly says:
        "Add aMethod with selector aSymbol to my
        methodDict. aMethod must not be defined locally."

but its somehow related with traits logic..
in case when you removing method from trait, and class has an original  
method,
it is restored back? no?



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

Comment #13 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

Sorry Igor I have no idea how it works. I am just telling you what I can  
see by putting a halt in #dropSourcePointer
If you want to try it yourself, just put a breakpoint in such method and  
evaluate:


Gofer new
        squeaksource: 'Fuel';
        package: 'Fuel';
load




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

Updates:
        Cc: [hidden email]

Comment #14 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

maybe Adrian can help with traits


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

Updates:
        Labels: -Milestone-1.3

Comment #15 on issue 4200 by [hidden email]: Build of Pharo Full  
does not terminate
http://code.google.com/p/pharo/issues/detail?id=4200

I removed the dropSourcePointer call for now. (13198)


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

Comment #16 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

Woith latest PharoCore 13198 which has the dropSourcePointer rollbacked,  
that is, no senders...the Compiler recompileAll stil takes a live.  
Impossible to do.



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

Comment #17 on issue 4200 by [hidden email]: Build of Pharo Full  
does not terminate
http://code.google.com/p/pharo/issues/detail?id=4200


[Compiler recompileAll] timeToRun 35019

with Cog...


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

Comment #18 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

marcus, are you sure you are trying in 13198?


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

Comment #19 on issue 4200 by marianopeck: Build of Pharo Full does not  
terminate
http://code.google.com/p/pharo/issues/detail?id=4200

I get a "computation has been terminated"  with cog


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