Re: [PATCH] Support --docdir, complain if that conflicts with
On Tuesday, 28 July 2009 at 10:57, Matthias Andree wrote:
> Am 28.07.2009, 05:51 Uhr, schrieb Brendan Cully <brendan@xxxxxxxxxx>:
>
> >On Monday, 27 July 2009 at 22:22, Matthias Andree wrote:
> >># HG changeset patch
> >># User Matthias Andree <matthias.andree@xxxxxx>
> >># Date 1248733349 -7200
> >># Branch HEAD
> >># Node ID 9949607b3f3d35121459cb34416ba3c1e54599bb
> >># Parent 86e066a206ac9d0eb2df5ce67eed22fda4287846
> >>Support --docdir, complain if that conflicts with --with-docdir.
> >>
> >>Fixes #3282.
> >>
> >>Note that this relies on the autoconf 2.60+ default for
> >>docdir; if that ever changes in a newer autoconf release, the script
> >>must be updated to avoid false positives.
> >>
> >>Note that --with-docdir should be phased out and replaced by --docdir,
> >>after 1.6.
> >
> >You're right, this is too fragile. I've pushed an alternative that
> >doesn't bother with the warning. If you specify both --docdir and
> >work, and the autoconf code is simple. I don't actually have a
> >pre-2.60 version to test with though.
>
> Sounds reasonable.
>
> But why not go all the way, require autoconf 2.60+ through AC_PREREQ
> and ditch --with-docdir? mutt 1.6 will be substantially different
> from mutt 1.4, so users expect things to be different. Mentioning in
> docs and release announcements that with-docdir is gone incompatibly
> would do the job. I don't see whose setup you'd be breaking with
> AC_PREREQ.
Bsically (although I'm happy to support it) I don't think
--with-docdir is a particularly valuable feature, so I'd rather not
take the chance that it causes some of the people who have been
tracking the repo to stop now, when testing is particularly valuable,
because for whatever reason their environment isn't completely up to
date.