[Pharo-project] [Fwd: [squeak-dev] Re: Bug in object finalization?]
So, now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream? Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox. afterwards a WeakRegistry default initialize: WeakRegistry default size will be necessary. Norbert
Hi Norbert, We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream. Adrian On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :) Have fun, Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Hi norbert did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :) Stef On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Thanks, Norbert, I integrated your change. I changed the migration script to make it also work if people do updates in an image that already has values registered for finalization, and for all existing weak registries, not only for the finalization registry. Like this the code and the state are in sync: WeakRegistry allInstancesDo: [ :each | (each instVarNamed: 'valueDictionary') primitiveChangeClassTo: WeakIdentityKeyDictionary basicNew; rehash ]. Cheers, Adrian On Jun 18, 2008, at 21:40 , Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Hi Stef, I harvested most of the remaining changes in the inbox, including Norberts fix (see other mail in this thread). - changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg I pushed everything into the Pharo repository, updated ScriptLoader (update10042), and moved the relevant packages from PharoInbox to PharoTreatedInbox. Cheer, Adrian On Jun 18, 2008, at 22:47 , Stéphane Ducasse wrote:
Hi norbert
did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :)
Stef
On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Wed, 2008-06-18 at 23:00 +0200, Adrian Lienhard wrote:
Thanks, Norbert, I integrated your change. I changed the migration script to make it also work if people do updates in an image that already has values registered for finalization, and for all existing weak registries, not only for the finalization registry. Like this the code and the state are in sync:
WeakRegistry allInstancesDo: [ :each | (each instVarNamed: 'valueDictionary') primitiveChangeClassTo: WeakIdentityKeyDictionary basicNew; rehash ].
Great! I would have done something similar if I only had the knowledge :) Norbert
Cheers, Adrian
On Jun 18, 2008, at 21:40 , Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Wed, 2008-06-18 at 22:47 +0200, Stéphane Ducasse wrote:
Hi norbert
did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :)
Yes, see my other mails. Everything is best now. Norbert
Stef
On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Wed, 2008-06-18 at 23:04 +0200, Adrian Lienhard wrote:
Hi Stef,
I harvested most of the remaining changes in the inbox, including Norberts fix (see other mail in this thread).
- changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg
I pushed everything into the Pharo repository, updated ScriptLoader (update10042), and moved the relevant packages from PharoInbox to PharoTreatedInbox.
I don't get it into the image via update stream. Is there an additional action to be taken for this to happen? Norbert
Cheer, Adrian
On Jun 18, 2008, at 22:47 , Stéphane Ducasse wrote:
Hi norbert
did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :)
Stef
On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Jun 18, 2008, at 23:19 , Norbert Hartl wrote:
On Wed, 2008-06-18 at 23:04 +0200, Adrian Lienhard wrote:
I don't get it into the image via update stream. Is there an additional action to be taken for this to happen?
yes, Stef usually pushes updates into the stream. Adrian
Norbert
Cheer, Adrian
On Jun 18, 2008, at 22:47 , Stéphane Ducasse wrote:
Hi norbert
did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :)
Stef
On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo- project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
excellent I will push the cs in the stream Stef On Jun 18, 2008, at 11:04 PM, Adrian Lienhard wrote:
Hi Stef,
I harvested most of the remaining changes in the inbox, including Norberts fix (see other mail in this thread).
- changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg
I pushed everything into the Pharo repository, updated ScriptLoader (update10042), and moved the relevant packages from PharoInbox to PharoTreatedInbox.
Cheer, Adrian
On Jun 18, 2008, at 22:47 , Stéphane Ducasse wrote:
Hi norbert
did you read the answer of nicolas on squeak-dev on your question. I read the mantis entry he mentioned but i read superficially. :)
Stef
On Jun 18, 2008, at 9:40 PM, Norbert Hartl wrote:
On Wed, 2008-06-18 at 21:37 +0200, Adrian Lienhard wrote:
Hi Norbert,
We do this kind of scripts in the ScriptLoader. Just let me know when you published your slice and I will prepare the ScriptLoader update script which Stef can then push into the update stream.
I published the slice 1 minute before sending the mail :)
Have fun,
Norbert
Adrian
On Jun 18, 2008, at 19:30 , Norbert Hartl wrote:
So,
now I'm all for fixing it and ressurect the test case. I'll upload something to PharoInbox. Afterwards it needs to be reinitialized. What is the preferred way of having doits in the update stream?
Anyway I uploaded SLICE-WeakRegistryAndFinalizerTestFix to PharoInbox.
afterwards a
WeakRegistry default initialize: WeakRegistry default size
will be necessary.
Norbert
From: Andreas Raab <andreas.raab@gmx.de> Date: June 18, 2008 7:02:11 PM GMT+02:00 To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Subject: [squeak-dev] Re: Bug in object finalization? Reply-To: The general-purpose Squeak developers list <squeak-dev@lists.squeakfoundation.org
Norbert Hartl wrote:
On Wed, 2008-06-18 at 09:22 -0700, Andreas Raab wrote:
Norbert Hartl wrote:
There was a test in 3.9 testing this (it failed). In 3.10 the test got removed but I think it is a bug.
What do you think? The test itself is buggy (see http://bugs.squeak.org/view.php?id=6347) but once you fix the test it illustrates a problem in WeakRegistry which needs to use a WeakIdentityKeyDictionary instead of a WeakKeyDictionary.
Ok, what I said. Do you see an harmful side effects by exchanging the dictionary?
No - we've fixed this a long time ago in our code base.
Cheers, - Andreas
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo- project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Hi Adrian I pushed the updates. Can you check if this is ok? Stef On Jun 18, 2008, at 11:04 PM, Adrian Lienhard wrote:
- changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg
adrian should I used al.448 which is in the inbox? Stef On Jun 18, 2008, at 11:04 PM, Adrian Lienhard wrote:
- changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg
On Thu, 2008-06-19 at 13:51 +0200, Stéphane Ducasse wrote:
Hi Adrian
I pushed the updates. Can you check if this is ok?
For me it is not working. It doesn't load any package from script132. I can't see what is going wrong but I checked for a few packages that are in script132 and they didn't get updated. Norbert
Yes this was also my impression. I will check 448 Stef On Jun 19, 2008, at 2:22 PM, Norbert Hartl wrote:
On Thu, 2008-06-19 at 13:51 +0200, Stéphane Ducasse wrote:
Hi Adrian
I pushed the updates. Can you check if this is ok?
For me it is not working. It doesn't load any package from script132. I can't see what is going wrong but I checked for a few packages that are in script132 and they didn't get updated.
Norbert
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
sorry my mistake I did a typo. stef On Jun 19, 2008, at 2:22 PM, Norbert Hartl wrote:
On Thu, 2008-06-19 at 13:51 +0200, Stéphane Ducasse wrote:
Hi Adrian
I pushed the updates. Can you check if this is ok?
For me it is not working. It doesn't load any package from script132. I can't see what is going wrong but I checked for a few packages that are in script132 and they didn't get updated.
Norbert
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Ah, my bad. I committed the improved script to the inbox. If you can still change it, take al.448, else it should not be a problem. ScriptLoader.al.448 just has a slightly better migration script. Adrian On Jun 19, 2008, at 13:53 , Stéphane Ducasse wrote:
adrian
should I used al.448 which is in the inbox?
Stef
On Jun 18, 2008, at 11:04 PM, Adrian Lienhard wrote:
- changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Stef, looking at the update stream it seems like you only need to copy ScriptLoader-al.448.mcz into the main repository. Adrian On Jun 19, 2008, at 14:31 , Stéphane Ducasse wrote:
sorry my mistake I did a typo.
stef
On Jun 19, 2008, at 2:22 PM, Norbert Hartl wrote:
On Thu, 2008-06-19 at 13:51 +0200, Stéphane Ducasse wrote:
Hi Adrian
I pushed the updates. Can you check if this is ok?
For me it is not working. It doesn't load any package from script132. I can't see what is going wrong but I checked for a few packages that are in script132 and they didn't get updated.
Norbert
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
done I created the update and I did a release Adrian this was great that you took the lead on that. please consider doing that more often :) Stef On Jun 19, 2008, at 2:38 PM, Adrian Lienhard wrote:
Stef, looking at the update stream it seems like you only need to copy ScriptLoader-al.448.mcz into the main repository. Adrian
On Jun 19, 2008, at 14:31 , Stéphane Ducasse wrote:
sorry my mistake I did a typo.
stef
On Jun 19, 2008, at 2:22 PM, Norbert Hartl wrote:
On Thu, 2008-06-19 at 13:51 +0200, Stéphane Ducasse wrote:
Hi Adrian
I pushed the updates. Can you check if this is ok?
For me it is not working. It doesn't load any package from script132. I can't see what is going wrong but I checked for a few packages that are in script132 and they didn't get updated.
Norbert
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
participants (3)
-
Adrian Lienhard -
Norbert Hartl -
Stéphane Ducasse