�Just remember to add corresponding deregistration in shutdown if SystemChangeNotifier is reimplemented using Announcements :)

(hint: Announcers do not care if objects register multiple times to the same Announcement)

Cheers,
Henry

On 01.09.2010 22:41, St�phane Ducasse wrote:

muchas gratias.


Stef

On Sep 1, 2010, at 9:08 PM, Adrian Lienhard wrote:

I created a changeset and attached it to the appropriate ticket: http://code.google.com/p/pharo/issues/detail?id=2896

Adrian

On Sep 1, 2010, at 09:52 , St�phane Ducasse wrote:

seems like we should do something :)

Begin forwarded message:

From: John M McIntosh<johnmci@smalltalkconsulting.com>
Date: September 1, 2010 1:12:28 AM GMT+02:00
To: St�phane Ducasse<stephane.ducasse@inria.fr>
Subject: Fwd: [squeak-dev] The Trunk: System-eem.366.mcz
Reply-To: johnmci@smalltalkconsulting.com

In case this gets overlooked being hidden in a update for The Trunk: System-eem.

The EndianCache does need to be cleared at shutdown versus in startup.
I can't speak for the SystemChangeNotifier since I'm not sure what the bug is..

Of course I'm not quite sure if anyone in the Pharo group looks and the trunk updates and moves things across?


Begin forwarded message:

From: commits@source.squeak.org
Date: August 31, 2010 5:47:35 PM PDT
To: squeak-dev@lists.squeakfoundation.org, packages@lists.squeakfoundation.org
Subject: [squeak-dev] The Trunk: System-eem.366.mcz
Reply-To: squeak-dev@lists.squeakfoundation.org

Eliot Miranda uploaded a new version of System to project The Trunk:
http://source.squeak.org/trunk/System-eem.366.mcz

==================== Summary ====================

Name: System-eem.366
Author: eem
Time: 31 August 2010, 10:47:17.97 am
UUID: 2126e5d9-4d47-4cc1-921f-daefd935aedd
Ancestors: System-nice.365

Make sure system notifications for Smalltalk are up-to-date
on start-up.
Fix voiding of Smalltalk's endian cache which needs to be done
in a timely manner.

=============== Diff against System-nice.365 ===============

Item was changed:
----- Method: SmalltalkImage class>>startUp (in category 'class initialization') -----
startUp
+ � � � SystemChangeNotifier uniqueInstance notify: Smalltalk ofAllSystemChangesUsing: #event:!
- � � � "XXXX: This is broken. SmalltalkImage startUp happens quite late in the startup sequence; earlier startups may very well need the information about the endianness of the platform."
- � � � EndianCache := nil.
- !

Item was changed:
----- Method: SmalltalkImage>>processShutDownList: (in category 'snapshot and quit') -----
processShutDownList: quitting
+ � � � "Send #shutDown to each class that needs to wrap up before a snapshot.
+ � � � �Also void the endianness chace; �this can't safely be done on start-up because
+ � � � �Smalltalk is too late in the start-up sequence."
- � � � "Send #shutDown to each class that needs to wrap up before a snapshot."

+ � � � EndianCache := nil.
+ � � � self send: #shutDown: toClassesNamedIn: ShutDownList with: quitting!
- � � � self send: #shutDown: toClassesNamedIn: ShutDownList with: quitting.
- !

--
===========================================================================
John M. McIntosh<johnmci@smalltalkconsulting.com> � �Twitter: �squeaker68882
Corporate Smalltalk Consulting Ltd. �http://www.smalltalkconsulting.com
===========================================================================

_______________________________________________
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