[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Sc-devel] SC3.2 RC 3 and SC 3.2 [was Re: APF fix]

2008/2/1, Josh Parmenter <josh@xxxxxxxxxxxxxxxxx>:
> On Feb 1, 2008, at 8:11 AM, Dan Stowell wrote:
> >
> >> Then, I think SC 3.2 should be out early next week (Tuesday or
> >> Wednesday).
> >
> > No. The release candidate needs to be out for long enough for users to
> > actually use the thing. Leave it at least a week after 3.2rc3 is out
> > before we declare it the official 3.2. In normal circumstances a week
> > would be a pathetically short time to have a release candidate out in
> > the wild, but in our context here we're never going to manage anything
> > better than that.
> >
> The main reason I think only a few days is needed is because there
> haven't really been any significant changes. I see this as a bugfix
> up, that mostly needs to be downloaded and ran. If users can do that,
> I don't think there is anything else that really needs testing...
> however, I was just suggesting times. We can space them out as much as
> anyone else wants (no problem to me).

The point of the "release candidate" approach is that projects like SC
are extremely complex systems, and there may be subtle things about
our bugfixes that we don't realise, but which end up f'ing stuff up
for users. If it was just a question of checking that it could be
downloaded and ran, there'd be no point doing RCs, we'd just upload
the files, download them once to check, and announce the release.

Sorry to bang on about this, but I think we need to invest some effort
in making *sure* that 3.2 is a really nice release, because of the
book connection in particular. The SC 3.1 release was untidy, since
everyone noticed problems *after* the release and ended up producing
3.1.1 a few days later.