Re: [Pharo-project] scriptloader>>loadComanche
I believe that the change is a proper one, so I fixed my code rather than report a bug:) It is related to the the asNumber changes discussion. In Metacello I was using asNumber to parse versionStrings and on an Error I resumed with the string itself instead of the number. In 10470 that technique worked and in 10487 I got an IllegalResumeAttempt ...Looking closer I see that asNumber now raises an Error and it didn't before. So the change was in how asNumber behaves, not in how Errors behave...sorry for the FUD. Dale ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote: | Hi dale | | this is strange because I do not remember us integrating a changes | that would do that. | Can you confirm it? | Stef | | | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | | > Johan, | > | > Yes there was a recent change to Error handling in Pharo that broke | | > Metacello (Error no longer resumable) that broke Metacello... I ran | | > into that one yesterday on 10487 and updated the Seaside Metacello | | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or later... | > | > Naturally as I try to verify that there aren't any other problems | > I'm running into network problems:) | > | > Dale | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | | > | >> We clearly need a universe like tools so that people can | publish | > | and | > | >> maintain their add-ons. | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | "flagship" | > | > until we have the infrastructur set up. | > | > | > | > Maybe we should delete it from the core but we can the loading | > | support | > | > in the dev images. What do you all think? | > | | > | I just loaded metacello and used its seaside configuration to load | | > the | > | | > | 3.0-alpha5 of seaside. | > | | > | It hits the 'reading a number failed' bug in 10487-DEV | > | | > | ...argh... | > | | > | | > | | > | ---------------------------- | > | Johan Brichau | > | johan.brichau@uclouvain.be | > | | > | | > | | > | | > | | > | _______________________________________________ | > | 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 Tue, Oct 20, 2009 at 2:50 PM, Dale Henrichs <dale.henrichs@gemstone.com>wrote:
I believe that the change is a proper one, so I fixed my code rather than report a bug:)
It is related to the the asNumber changes discussion. In Metacello I was using asNumber to parse versionStrings and on an Error I resumed with the string itself instead of the number. In 10470 that technique worked and in 10487 I got an IllegalResumeAttempt ...Looking closer I see that asNumber now raises an Error and it didn't before.
So the change was in how asNumber behaves, not in how Errors behave...sorry for the FUD.
Yes, that was what we were discussing in the thread with subject '' asNumber -> 'Reading a number failed' I understood it was rollbacked. Mariano
Dale ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote:
| Hi dale | | this is strange because I do not remember us integrating a changes | that would do that. | Can you confirm it? | Stef | | | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | | > Johan, | > | > Yes there was a recent change to Error handling in Pharo that broke | | > Metacello (Error no longer resumable) that broke Metacello... I ran | | > into that one yesterday on 10487 and updated the Seaside Metacello | | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or later... | > | > Naturally as I try to verify that there aren't any other problems | > I'm running into network problems:) | > | > Dale | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | | > | >> We clearly need a universe like tools so that people can | publish | > | and | > | >> maintain their add-ons. | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | "flagship" | > | > until we have the infrastructur set up. | > | > | > | > Maybe we should delete it from the core but we can the loading | > | support | > | > in the dev images. What do you all think? | > | | > | I just loaded metacello and used its seaside configuration to load | | > the | > | | > | 3.0-alpha5 of seaside. | > | | > | It hits the 'reading a number failed' bug in 10487-DEV | > | | > | ...argh... | > | | > | | > | | > | ---------------------------- | > | Johan Brichau | > | johan.brichau@uclouvain.be | > | | > | | > | | > | | > | | > | _______________________________________________ | > | 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
That's good, because I just ran into a place where asNumber wasn't protected by a handler ... I haven't tried this on the latest Pharo. but I'll give it a try now. Thanks, Dale ----- "Mariano Martinez Peck" <marianopeck@gmail.com> wrote: | On Tue, Oct 20, 2009 at 2:50 PM, Dale Henrichs | <dale.henrichs@gemstone.com>wrote: | | > I believe that the change is a proper one, so I fixed my code rather | than | > report a bug:) | > | > It is related to the the asNumber changes discussion. In Metacello | I was | > using asNumber to parse versionStrings and on an Error I resumed | with the | > string itself instead of the number. In 10470 that technique worked | and in | > 10487 I got an IllegalResumeAttempt ...Looking closer I see that | asNumber | > now raises an Error and it didn't before. | > | > So the change was in how asNumber behaves, not in how Errors | behave...sorry | > for the FUD. | > | > | Yes, that was what we were discussing in the thread with subject '' | asNumber -> 'Reading a number failed' | | I understood it was rollbacked. | | Mariano | | | > Dale | > ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote: | > | > | Hi dale | > | | > | this is strange because I do not remember us integrating a | changes | > | that would do that. | > | Can you confirm it? | > | Stef | > | | > | | > | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | > | | > | > Johan, | > | > | > | > Yes there was a recent change to Error handling in Pharo that | broke | > | | > | > Metacello (Error no longer resumable) that broke Metacello... I | ran | > | | > | > into that one yesterday on 10487 and updated the Seaside | Metacello | > | | > | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or | later... | > | > | > | > Naturally as I try to verify that there aren't any other | problems | > | > I'm running into network problems:) | > | > | > | > Dale | > | > | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | > | | > | > | >> We clearly need a universe like tools so that people can | > | publish | > | > | and | > | > | >> maintain their add-ons. | > | > | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | > | "flagship" | > | > | > until we have the infrastructur set up. | > | > | > | > | > | > Maybe we should delete it from the core but we can the | loading | > | > | support | > | > | > in the dev images. What do you all think? | > | > | | > | > | I just loaded metacello and used its seaside configuration to | load | > | | > | > the | > | > | | > | > | 3.0-alpha5 of seaside. | > | > | | > | > | It hits the 'reading a number failed' bug in 10487-DEV | > | > | | > | > | ...argh... | > | > | | > | > | | > | > | | > | > | ---------------------------- | > | > | Johan Brichau | > | > | johan.brichau@uclouvain.be | > | > | | > | > | | > | > | | > | > | | > | > | | > | > | _______________________________________________ | > | > | 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 Tue, Oct 20, 2009 at 3:04 PM, Dale Henrichs <dale.henrichs@gemstone.com>wrote:
That's good, because I just ran into a place where asNumber wasn't protected by a handler ... I haven't tried this on the latest Pharo. but I'll give it a try now.
The same happens to me with the port of Glorp to Pharo :( Do a system update before because this change was after the latest image that was generated and uploaded.
Thanks,
Dale
----- "Mariano Martinez Peck" <marianopeck@gmail.com> wrote:
| On Tue, Oct 20, 2009 at 2:50 PM, Dale Henrichs | <dale.henrichs@gemstone.com>wrote: | | > I believe that the change is a proper one, so I fixed my code rather | than | > report a bug:) | > | > It is related to the the asNumber changes discussion. In Metacello | I was | > using asNumber to parse versionStrings and on an Error I resumed | with the | > string itself instead of the number. In 10470 that technique worked | and in | > 10487 I got an IllegalResumeAttempt ...Looking closer I see that | asNumber | > now raises an Error and it didn't before. | > | > So the change was in how asNumber behaves, not in how Errors | behave...sorry | > for the FUD. | > | > | Yes, that was what we were discussing in the thread with subject '' | asNumber -> 'Reading a number failed' | | I understood it was rollbacked. | | Mariano | | | > Dale | > ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote: | > | > | Hi dale | > | | > | this is strange because I do not remember us integrating a | changes | > | that would do that. | > | Can you confirm it? | > | Stef | > | | > | | > | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | > | | > | > Johan, | > | > | > | > Yes there was a recent change to Error handling in Pharo that | broke | > | | > | > Metacello (Error no longer resumable) that broke Metacello... I | ran | > | | > | > into that one yesterday on 10487 and updated the Seaside | Metacello | > | | > | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or | later... | > | > | > | > Naturally as I try to verify that there aren't any other | problems | > | > I'm running into network problems:) | > | > | > | > Dale | > | > | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | > | | > | > | >> We clearly need a universe like tools so that people can | > | publish | > | > | and | > | > | >> maintain their add-ons. | > | > | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | > | "flagship" | > | > | > until we have the infrastructur set up. | > | > | > | > | > | > Maybe we should delete it from the core but we can the | loading | > | > | support | > | > | > in the dev images. What do you all think? | > | > | | > | > | I just loaded metacello and used its seaside configuration to | load | > | | > | > the | > | > | | > | > | 3.0-alpha5 of seaside. | > | > | | > | > | It hits the 'reading a number failed' bug in 10487-DEV | > | > | | > | > | ...argh... | > | > | | > | > | | > | > | | > | > | ---------------------------- | > | > | Johan Brichau | > | > | johan.brichau@uclouvain.be | > | > | | > | > | | > | > | | > | > | | > | > | | > | > | _______________________________________________ | > | > | 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 | >
Mariano, The change hasn't been pushed all the way out yet. GemStone errors out on invalid input to asNumber (but GemStone Errors are resumable ... which is why I fixed the error handler) so I will fix the other case that I found in Metacello, anyway. Dale ----- "Mariano Martinez Peck" <marianopeck@gmail.com> wrote: | On Tue, Oct 20, 2009 at 2:50 PM, Dale Henrichs | <dale.henrichs@gemstone.com>wrote: | | > I believe that the change is a proper one, so I fixed my code rather | than | > report a bug:) | > | > It is related to the the asNumber changes discussion. In Metacello | I was | > using asNumber to parse versionStrings and on an Error I resumed | with the | > string itself instead of the number. In 10470 that technique worked | and in | > 10487 I got an IllegalResumeAttempt ...Looking closer I see that | asNumber | > now raises an Error and it didn't before. | > | > So the change was in how asNumber behaves, not in how Errors | behave...sorry | > for the FUD. | > | > | Yes, that was what we were discussing in the thread with subject '' | asNumber -> 'Reading a number failed' | | I understood it was rollbacked. | | Mariano | | | > Dale | > ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote: | > | > | Hi dale | > | | > | this is strange because I do not remember us integrating a | changes | > | that would do that. | > | Can you confirm it? | > | Stef | > | | > | | > | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | > | | > | > Johan, | > | > | > | > Yes there was a recent change to Error handling in Pharo that | broke | > | | > | > Metacello (Error no longer resumable) that broke Metacello... I | ran | > | | > | > into that one yesterday on 10487 and updated the Seaside | Metacello | > | | > | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or | later... | > | > | > | > Naturally as I try to verify that there aren't any other | problems | > | > I'm running into network problems:) | > | > | > | > Dale | > | > | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | > | | > | > | >> We clearly need a universe like tools so that people can | > | publish | > | > | and | > | > | >> maintain their add-ons. | > | > | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | > | "flagship" | > | > | > until we have the infrastructur set up. | > | > | > | > | > | > Maybe we should delete it from the core but we can the | loading | > | > | support | > | > | > in the dev images. What do you all think? | > | > | | > | > | I just loaded metacello and used its seaside configuration to | load | > | | > | > the | > | > | | > | > | 3.0-alpha5 of seaside. | > | > | | > | > | It hits the 'reading a number failed' bug in 10487-DEV | > | > | | > | > | ...argh... | > | > | | > | > | | > | > | | > | > | ---------------------------- | > | > | Johan Brichau | > | > | johan.brichau@uclouvain.be | > | > | | > | > | | > | > | | > | > | | > | > | | > | > | _______________________________________________ | > | > | 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 Oct 20, 2009, at 7:50 PM, Dale Henrichs wrote:
I believe that the change is a proper one, so I fixed my code rather than report a bug:)
It is related to the the asNumber changes discussion. In Metacello I was using asNumber to parse versionStrings and on an Error I resumed with the string itself instead of the number. In 10470 that technique worked and in 10487 I got an IllegalResumeAttempt ...Looking closer I see that asNumber now raises an Error and it didn't before.
So the change was in how asNumber behaves, not in how Errors behave...sorry for the FUD.
no problem. Now we had a discussion and: - adrian rollbacked the changes for 1.0 - we got it in 1.1 - we will probably propose two methods one for parsing number and the other trying to parse anything having number in it.
Dale ----- "Stéphane Ducasse" <stephane.ducasse@inria.fr> wrote:
| Hi dale | | this is strange because I do not remember us integrating a changes | that would do that. | Can you confirm it? | Stef | | | On Oct 20, 2009, at 7:29 PM, Dale Henrichs wrote: | | > Johan, | > | > Yes there was a recent change to Error handling in Pharo that broke | | > Metacello (Error no longer resumable) that broke Metacello... I ran | | > into that one yesterday on 10487 and updated the Seaside Metacello | | > config (Seaside-Metacello-Configuration-dkh.15) with a fix. | > | > For Pharo 10487 you'll need to use Metacello 1.0-beta.6 or later... | > | > Naturally as I try to verify that there aren't any other problems | > I'm running into network problems:) | > | > Dale | > | > ----- "Johan Brichau" <johan.brichau@uclouvain.be> wrote: | > | > | On 20 Oct 2009, at 12:00, Torsten Bergmann wrote: | > | | > | >> We clearly need a universe like tools so that people can | publish | > | and | > | >> maintain their add-ons. | > | > | > | > Yes, but we should be pragmatic for Seaside since it is the | > | "flagship" | > | > until we have the infrastructur set up. | > | > | > | > Maybe we should delete it from the core but we can the loading | > | support | > | > in the dev images. What do you all think? | > | | > | I just loaded metacello and used its seaside configuration to load | | > the | > | | > | 3.0-alpha5 of seaside. | > | | > | It hits the 'reading a number failed' bug in 10487-DEV | > | | > | ...argh... | > | | > | | > | | > | ---------------------------- | > | Johan Brichau | > | johan.brichau@uclouvain.be | > | | > | | > | | > | | > | | > | _______________________________________________ | > | 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)
-
Dale Henrichs -
Mariano Martinez Peck -
Stéphane Ducasse