Hi Edgar,

On Fri, Mar 30, 2012 at 5:51 AM, Bert Freudenberg <bert at freudenbergs.de>wrote:

>
> On 30.03.2012, at 13:39, Edgar J. De Cleene wrote:
>
> >
> > ------ Forwarded Message
> >> From: <commits at source.squeak.org>
> >> Reply-To: <squeak-dev at lists.squeakfoundation.org>
> >> Date: Fri, 30 Mar 2012 10:56:29.998 0000
> >> To: <squeak-dev at lists.squeakfoundation.org>,
> >> <packages at lists.squeakfoundation.org>
> >> Subject: [squeak-dev] The Trunk: System-ul.477.mcz
> >>
> >> Levente Uzonyi uploaded a new version of System to project The Trunk:
> >> http://source.squeak.org/trunk/System-ul.477.mcz
> >>
> >> ==================== Summary ====================
> >>
> >> Name: System-ul.477
> >> Author: ul
> >> Time: 30 March 2012, 12:27:45.894 pm
> >> UUID: fb2ac551-9367-134f-9872-2f65037f69ff
> >> Ancestors: System-dtl.476
> >>
> >> - restored the original behavior of SystemNavigation >>
> #allCallsOn:from:, so
> >> it returns calls in subclasses too
> >> - temporarily restored NaturalLanguageTranslator class >>
> #translationSuffix
> >> so FileList can be used again
> >>
> >> =============== Diff against System-dtl.476 ===============
> >>
> >> Item was added:
> >> + ----- Method: NaturalLanguageTranslator class>>translationSuffix (in
> >> category 'accessing') -----
> >> + translationSuffix
> >> +
> >> +  ^'translation'!
> >>
> >> Item was changed:
> >>  ----- Method: SystemNavigation>>allCallsOn:from: (in category 'query')
> -----
> >>  allCallsOn: aSymbol from: aBehavior
> >> +  "Answer a sorted collection of all the methods from aBehavior and
> it's
> >> subclasses that call on aSymbol."
> >> -  "Answer a sorted collection of all the methods from aBehavior that
> call on
> >> aSymbol."
> >>
> >> +  ^self allCallsOn: aSymbol fromBehaviors: aBehavior withAllSubclasses
> >> sorted: true!
> >> -  ^self allCallsOn: aSymbol fromBehaviors: { aBehavior } sorted: true!
> >
> > ------ End of Forwarded Message
> >
> >
> > I working....
> >
> > Maybe not enough smart or not enough fast.
> >
> > But I found the trouble Nicolas report and delete all file-services in
> > NaturalLanguageTranslator
> > And do FileList initialize
> >
> > Save the Squeak4.4-11969.image and when restarting no problem.
> >
> > Figuring how traslate this to trunk as "we don't like have working
> .image in
> > ftp"
> >
> > Maybe I should learn to have my mouth closed and do not volunteering.
> > But Squeak is in comatose state.
> >
> > If my mistakes awake some Masters, then so be it
> >
> > Regards
> >
> > Edgar
>
>
>
> Not everyone can download a new image you made by hand. We have our
> working images with all kinds of stuff in it, and we want to update those
> images, not switching to new ones. Also, the package history serves as a
> record of what changed. That's why it is important to fix the stuff in the
> Trunk packages, and not fiddle with images. The images on the ftp server
> are pretty much only for release, and release test candidates.
>
> - Bert -
>

I just wanted to second what Bert is saying. Images are wonderful because
one can tailor them, but they can become a prison from which your code can
never escape, and from which you can never move forward, *unless* they can
be updated automatically.  Note that the Monticello update process does a
merge by default so you can even update modified packages to a certain
degree.  So keeping the update process working is *really, really* valuable
and hence *really* worth the effort.

If (as we should) we all have our own tailored images then we can't expect
anyone else to use our images, because we would be inflicting our tailoring
on them and asking them to give up the tailoring they had done.  We all
have to be free to tailor in our own way.  So forcing people to download a
non-release image to update code is simply unacceptable.

Courage, and thanks so much for your efforts!
-- 
best,
Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20120331/aba8ce57/attachment.htm