Re: libtool release schedule [WAS: [PATCH] GNU/KNetBSD support]

[Gary V. Vaughan]

Robert Millan wrote:
> On Tue, Sep 23, 2003 at 12:05:43PM +0100, Gary V. Vaughan wrote:
>
> > > On Mon, Sep 22, 2003 at 01:22:42PM +0100, Gary V. Vaughan wrote:
> > >
> > >
> > > > Gotta reinstate my shell account at the FSF,
>
>
> I've just spoken with Jeff Bailey (from the admin team) on that. The issue
> was on hold for a script that automated gpg-signed FTP uploads. Yesterday I
> wrote such script and sent it to him. Expect to recover access soon!

Cool, thanks!

> > I'd rather polish 1.6.  But if 1.6 is essentially done (i.e branched in CVS 
> > and 1.5b released to alpha.gnu.org), I'd be prepared to roll a 1.5.1 
> > release if we are simply waiting on automake 1.8.
>
>
> What do you need for 1.5b? Automake 1.8 too?

The tasks I listed earlier need to be complete, automake-1.8 needs to be
released since libtool HEAD now relies on it, we need to make sure the test
suite passes on as many platforms as possible, and fix it for those that have
regressions.  The mechanics of the release are documented in README-alpha in
CVS. (I've noticed that the web update step, and md5sum generation are missing
though).

> > Unfortunately the full release process is a fairly time consuming process, 
> > so I don't want to spend an evening rolling 1.5.1 if 1.6 is only a month 
> > behind if I have other stuff to work on.   But I do take your point.
>
> Is there anything else from the release process you can delegate me? I'm not
> aware of the required tasks asides from the file upload itself.

There is a bunch of CVS and website administration associated with a release,
and also the announcement needs to be compiled and sent to various mailing 
lists.

> For example if you want testing on exotic platforms, I can try branch-1-5
> and/or HEAD on most of them for you. For CPUs I have access to Debian machines
> and can test it on m68k, sparc[64], alpha, powerpc, arm, mips (l.e. and b.e.),
> hppa, ia64, s390, sh and x86_64. For OSes I can confirm GNU/Hurd [1] and try
> a few others.

That would be an enormous help, especially if you can help track down any
problems we uncover.  Theoretically, the code churn on branch-1-5 should not
cause any regressions, but if you want to test the head of that branch and 
send updates for the PLATFORMS file, that would be okay.  If your time is more
limited then please wait until we make branch-1-6 and perform the tests with
the head of that branch after 1.5b is released from it.

Cheers,
        Gary.
--
  ())_.  Gary V. Vaughan    gary@(lilith.warpmail.net|gnu.org)
  ( '/   Research Scientist http://www.oranda.demon.co.uk       ,_())____
  / )=   GNU Hacker         http://www.gnu.org/software/libtool  \'      `&
`(_~)_   Tech' Author       http://sources.redhat.com/autobook   =`---d__/