2010/1/11 Miguel Enrique Cob� Martinez
<miguel.coba@gmail.com>
El lun, 11-01-2010 a las 22:06 +0000, Michael Roberts escribi�:
> I'm afraid i've got rather lost in this thread.
>
> >> �As an after thought after seeing other posts on the issue, I think
> >> �more �than �a �generic �warning �that "may �break �the �image" �the
> >> �Monticello comments and/or a �simple explanation on the reasons and
> >> �affected �packages *before* �the update �starts with �an �option to
> >> �proceed or abort is much in need.
> >>
> >>
> >> �Yes, maybe. �Then thing �is that I �am not �sure in which �cases it
> >> would break. Only �in overrides? How many overrides �do we have from
> >> external pacakges in Core stuff? �I have no idea.
> >
>
> I think given that we are at a very early stage of building the Pharo
> image from PharoCore i.e Metacello just coming on etc, we should
> simplify the concerns above.
>
> Basically we should only build Pharo on top of a known PharoCore. Just
> because packages don't contain overrides does not imply that moving
> the core image will break packages on top. Don't offer such a feature,
> and then worry about what will happen.
>
> Part of the Pharo build script should set a flag that prevents system
> update in the short term. Longer term we should switch the available
> update mechanism to #loadLatest on the ConfigurationOfPharo. �There is
> then one System update feature, but different implementation in Pharo
> vs PharoCore.
>
So the vote is for disabled for now?
Yes. And this is cool because we have to put focus in to make Pharo images more stable. We cannot introduce such a feature when we are not even stable with Dev images. Let's go step by step. Once we get stable Dev images, good processes and package managment, then we can review this stuff about updating.
Do you want to do it ?� ;)
�
--