Hannes Hirzel wrote:
Hi Doug
Doug Way dway@riskmetrics.com wrote:
Okay, I'm very easily talked out of delaying the 3.4 release any further :-) (and was having second thoughts after posting that last message), so I agree with the prevailing opinion to leave this change out and go ahead with the release.
Well of course you guides can do anything you want "releasing" things, i.e. putting a label on a jar and call it a "release". Not having a proper test culture in Squeak leads us to have this kind of funny discussions. Measuring risk just with a feeling in the guts? "I feel like releaseing... I'm scary of metaclasses ....people don't use them anyway...".
Well, the way I phrased it ("I'm very easily talked out of delaying the 3.4 release...") may have made it seem that I suddenly decided to go ahead with the release on a whim.
But that's not really the case. I (and the other Guides) believe that leaving this fix out is the right thing to do at this stage of the release. We do have some (admittedly rough) standards that we're trying to follow during the release cycle. See: http://swiki.squeakfoundation.org/squeakfoundation/89
3.4 is currently in "gamma" status, which is basically a candidate final release, and that means that only critical problems should hold up the release. It's hard to define "critical" precisely, but it would have to be something that made the system unusable to a large portion of users. The ClassBuilder problem is significant, but would really only affect people who need to rebuild all of their classes, which is only required by a few (relatively experimental) packages.
And there may be significant risk in adding this fix to 3.4 at this late stage... changing the ClassBuilder is a pretty fundamental change. It might have some subtle side effect on the way classes are made obsolete, I don't know. I'm sure Andreas would not absolutely guarantee that it is side-effect-free. :-) So, even allowing 3 days to test this sort of change might not be enough.
Anyway, these are typical issues for any software release. At some point you do have to decide that enough changes/fixes have gone into a particular release, and we need to put the remaining larger changes/fixes into the next release. I was starting to fall into the "just one more fix" trap myself here.
Of course there needs to be an overall plan, too, which we roughly adhered to for 3.4, although the final release date is now quite late for various reasons, mostly related to the Guides taking over responsibility from SqC. (We are new at this, after all.)
Regarding 3.5 content, I agree that that needs to be discussed... I imagine discussions will start up on that topic soon on the SqF list. (Ah, I see you just started a thread here on squeak-dev...)
- Doug Way
squeak-dev@lists.squeakfoundation.org