21.4 not tagged in CVS

21.4 not tagged in CVS

[Jason Rumney]

It appears that normal release procedures have been bypassed in
releasing 21.4.

CVS contains no tags for 21.4, but it appears to have been created
from EMACS_21_3 except for movemail.c.

This is unfortunate, as we almost released 21.4 a year ago to fix a
build problem with recent versions of GCC on Windows, and a few other
"annoying bugs" that are fixed on the EMACS_21_1_RC branch.

The discussion is here:

http://lists.gnu.org/archive/html/help-emacs-windows/2004-02/msg00028.html


Instead, I think I tracked down an older copy of gcc to compile the
release, and checked in the changes needed for recent versions of gcc
in case there was a future release from the branch. I upgraded gcc
again after that, and did not expect to need it again now that the bug
is fixed in all current branches of CVS. I was not expecting an Ad-Hoc
release to be made like this.

Re: 21.4 not tagged in CVS

[David Kastrup]

Jason Rumney <jasonr@gnu.org> writes:

> It appears that normal release procedures have been bypassed in
> releasing 21.4.
>
> CVS contains no tags for 21.4, but it appears to have been created
> from EMACS_21_3 except for movemail.c.
>
> This is unfortunate, as we almost released 21.4 a year ago to fix a
> build problem with recent versions of GCC on Windows, and a few other
> "annoying bugs" that are fixed on the EMACS_21_1_RC branch.
>
> The discussion is here:
>
> http://lists.gnu.org/archive/html/help-emacs-windows/2004-02/msg00028.html
>
> Instead, I think I tracked down an older copy of gcc to compile the
> release, and checked in the changes needed for recent versions of
> gcc in case there was a future release from the branch. I upgraded
> gcc again after that, and did not expect to need it again now that
> the bug is fixed in all current branches of CVS. I was not expecting
> an Ad-Hoc release to be made like this.

Nobody expected it.  It was a special situation that arose.  The way
Richard has described it, I don't think that we can at all rule out
that a similar situation might happen again.  The solution would seem
to have one branch where critical fixes are applied to regardless of
whether a release is anticipated.  I am not sure this would be
something that in general would warrant the effort.  This would have
to be estimated by those that are capable of actually rolling a
release.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: 21.4 not tagged in CVS

[Jason Rumney]

David Kastrup <dak@gnu.org> writes:

> Nobody expected it.  It was a special situation that arose.

I don't think that is a good excuse to ignore CVS and make an ad-hoc
release that is not labelled, nor as far as I can see not even checked
in to CVS.

> The way Richard has described it, I don't think that we can at all
> rule out that a similar situation might happen again.

Which is why it is even more of a concern that this release was so
ad-hoc. There is a danger now that if it happens again, we will lose
the movemail patches.

> The solution would seem to have one branch where critical fixes are
> applied to regardless of whether a release is anticipated.

EMACS_21_1_RC was supposed to be such a branch.

Re: 21.4 not tagged in CVS

[Richard Stallman]

At this point I don't think it is worth trying to make another
release from the "release branch".  We are headed for a release
from the trunk, and we can see it coming closer.  So if we have
to make a 21.5 release, it will be like the 21.4 release, and
we may as well handle it the same way.