Re: New pretest tarball 22.0.93

[Chong Yidong]

"Juanma Barranquero" <address@hidden> writes:

> On 1/26/07, Eli Zaretskii <address@hidden> wrote:
> 
> > 4 pretest tarballs is not too many, I think.
> 
> No, of course not. I'm more worried by the month-long interval between
> pretests than the number of them.

The main reason for this is that Richard has, in the past, insisted on
eliminating outstanding bugs before starting a new pretest.  So I have
waited for the items in FOR-RELEASE items to be depleted before
rolling a new tarball.  There is certainly no difficulty for me to
roll a tarball every week if we decide on that.

> No, of course not. I'm more worried by the month-long interval between
> pretests than the number of them. 30 days is enough for us to grow
> restless and leave us susceptible to featurettes posing as
> bug-fixes... (I have this mental image of us developers like the guys
> in Starship Troopers, surrounded by nasty, mean-looking bugs.)

If we were to release Emacs 22 right now, it would already be much
more polished than the average free software release.  Personally, I
think that at this point, getting 5 years of accumulated crash and
data-corruption bugfixes into users' hands is more important that
scraping through the remaining handful of niggly bugs left in the
code---that ought to be what 22.2 is for.  But we've been through this
discussion before.

> > during the pretest of Emacs 21,
> > I think many pretesters would send email short time after another
> > pretest was tarred, saying they have the new version up and running.
> > Perhaps we should ask them to provide this information now.
> 
> Agreed.

One problem is Richard's time-lag.  He has to receive my email through
emacs-devel about the pretest tarball, then send the email to the
pretesters list, which usually takes several days.  I could send out
the pretest email directly if that makes things easier.