[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] Shell selection bug fixes
From: |
Eric Sunshine |
Subject: |
Re: [PATCH] Shell selection bug fixes |
Date: |
Tue, 13 Jan 2004 11:17:06 -0500 |
On Tue, 13 Jan 2004 12:57:26 +0000, Patrick Welche wrote:
> On Wed, Jan 07, 2004 at 05:08:55PM -0500, Eric Sunshine wrote:
> > Here is a patch which corrects several problems with the way the Autoconf
> > CVS version of AS_SHELL_SANITIZE attempts to locate a usable shell. The
> > patch address the following issues.
> Is there any chance this patch could be applied, and the cvs files
> it affects updated? Without it I just get the traditional "Found
> no shell that has working shell functions." which isn't right
> on NetBSD-current/i386. With it, life is better, but I still get
> that message from tools (like autom4te) which where expanded from
> the old macros. I got stuck after a maintainer-clean trying to
> rebuild those files... a bit of a chicken and egg situtation.
We have been having a discussion on autoconf-patches about a different patch
submitted by Paolo Bonzini which also addresses this issue, and which
provides a more generic mechanism for performing feature tests. In the long
run, the generic mechanism will be more appropriate. Upon reviewing Paolo's
patch, however, I discovered a number of problems with it (some of which are
fatal); but we are hoping that he can amend his patch to correct these
problems, and that the patch can then be installed. Since Paolo's patch
provides a better solution for the future, I suggested that it would make
more sense to ignore my patch and go with Paolo's instead. At this point, we
are awaiting a revised patch from Paolo.
As an interim solution, can you install Autoconf 2.59 in order to get our of
your chicken-and-egg problem? Alternately, perhaps my patch can be
committed temporarily (and those generated files, such as autom4te.in, can
corrected); at least until Paolo's revised patch becomes available.
Unfortunately, it seems that the GNU mailing list software has sent our
entire discussion on autoconf-patches to /dev/null (as it so often does), so
it has not been seen by anyone not included via cc:. At first I thought that
the autoconf-patches list was simply down, but a message to the list in a
different thread did arrive on the list today, so I am now assuming that our
entire discussion is probably gone for good.
-- ES