headers
Alexandre Julliard | 1 Aug 2006 19:34
Favicon

Typo in AC_PATH_XTRA always sets ac_werror_flag

The latest change to AC_PATH_XTRA contains a typo that makes it always
set the ac_werror_flag to a non-empty value instead of restoring the
previous value. Here's the trivial fix:

2006-08-01  Alexandre Julliard  <julliard <at> winehq.org>

	* lib/autoconf/libs.m4 (AC_PATH_XTRA): Fixed a typo
	in the restoring of the werror flag.

Index: lib/autoconf/libs.m4
===================================================================
RCS file: /sources/autoconf/autoconf/lib/autoconf/libs.m4,v
retrieving revision 1.20
diff -u -p -r1.20 libs.m4
--- lib/autoconf/libs.m4	20 Jul 2006 22:38:47 -0000	1.20
+++ lib/autoconf/libs.m4	1 Aug 2006 17:16:34 -0000
 <at>  <at>  -409,7 +409,7  <at>  <at>  else
 	 [AC_MSG_RESULT([yes])
 	  X_LIBS="$X_LIBS -R $x_libraries"],
 	 [AC_MSG_RESULT([neither works])])])
-    ac_[]_AC_LANG_ABBREV[]_werror_flag=ac_xsave_[]_AC_LANG_ABBREV[]_werror_flag
+    ac_[]_AC_LANG_ABBREV[]_werror_flag=$ac_xsave_[]_AC_LANG_ABBREV[]_werror_flag
     LIBS=$ac_xsave_LIBS
   fi

--

-- 
Alexandre Julliard
julliard <at> winehq.org
(Continue reading)

Permalink | Reply | 1 comment |
headers
Patrick Welche | 3 Aug 2006 16:54
Picon
Picon

Re: AC_PATH_XTRA and -rpath

On Thu, Jul 20, 2006 at 07:03:27PM -0700, Russ Allbery wrote:
> Paul Eggert <eggert <at> CS.UCLA.EDU> writes:
> > "Peter O'Gorman" <peter <at> pogma.com> writes:
> 
> >> This patch will now add -R to X_LIBS for all systems.
> 
> > It's not supposed to.  It's supposed to try -R, and add it only if it
> > worked.
> 
> Adding -R for system X libraries is always the wrong thing to do on Debian
> systems and will break multiarch builds.  I really don't think this is the
> right solution.

How about something like

% objdump -x /bin/ls | grep RPATH
  RPATH       /lib

then use rpath ?

Cheers,

Patrick
Permalink | Reply | 0 comments |
headers
Russ Allbery | 4 Aug 2006 01:58
Picon
Favicon
Gravatar

Re: AC_PATH_XTRA and -rpath

Patrick Welche <prlw1 <at> newn.cam.ac.uk> writes:
> On Thu, Jul 20, 2006 at 07:03:27PM -0700, Russ Allbery wrote:

>> Adding -R for system X libraries is always the wrong thing to do on
>> Debian systems and will break multiarch builds.  I really don't think
>> this is the right solution.

> How about something like

> % objdump -x /bin/ls | grep RPATH
>   RPATH       /lib

> then use rpath ?

I don't understand what your proposal means.

Systems that set an RPATH pointing to /lib for things like ls are probably
going to have a devil of a time doing multilib properly.

--

-- 
Russ Allbery (rra <at> stanford.edu)             <http://www.eyrie.org/~eagle/>
Permalink | Reply | 0 comments |
headers
Harlan Stenn | 4 Aug 2006 02:18

Re: AC_PATH_XTRA and -rpath

I thought he was ssaying that if /bin/ls uses RPATH then configure
should use RPATH, too.

H
--
> Patrick Welche <prlw1 <at> newn.cam.ac.uk> writes:
> > On Thu, Jul 20, 2006 at 07:03:27PM -0700, Russ Allbery wrote:
> 
> >> Adding -R for system X libraries is always the wrong thing to do on
> >> Debian systems and will break multiarch builds.  I really don't think
> >> this is the right solution.
> 
> > How about something like
> 
> > % objdump -x /bin/ls | grep RPATH
> >   RPATH       /lib
> 
> > then use rpath ?
> 
> I don't understand what your proposal means.
> 
> Systems that set an RPATH pointing to /lib for things like ls are probably
> going to have a devil of a time doing multilib properly.
Permalink | Reply | 0 comments |
headers
Patrick Welche | 4 Aug 2006 16:00
Picon
Picon

Re: AC_PATH_XTRA and -rpath

On Fri, Aug 04, 2006 at 12:18:11AM +0000, Harlan Stenn wrote:
> I thought he was ssaying that if /bin/ls uses RPATH then configure
> should use RPATH, too.

That's right - the question was how to know whether or not a system
uses rpath without reinventing libtool. In fact choosing some binary
we know exists outside of /bin /sbin might be better, e.g., sed.

Cheers,

Patrick
Permalink | Reply | 0 comments |
headers
Paul Eggert | 6 Aug 2006 08:26
Favicon

Re: Typo in AC_PATH_XTRA always sets ac_werror_flag

Alexandre Julliard <julliard <at> winehq.org> writes:

> 2006-08-01  Alexandre Julliard  <julliard <at> winehq.org>
>
> 	* lib/autoconf/libs.m4 (AC_PATH_XTRA): Fixed a typo
> 	in the restoring of the werror flag.

Thanks; I installed that.
Permalink | Reply | 0 comments |
headers
Paul Eggert | 7 Aug 2006 07:40
Favicon

Re: [GNU Autoconf 2.60] testsuite: 3 failed

Thanks for the bug report.  You can apply the following patch:

http://lists.gnu.org/archive/html/autoconf-patches/2006-07/msg00028.html
Permalink | Reply | 0 comments |
headers
Paul Eggert | 7 Aug 2006 07:43
Favicon

any objections to Autoconf 2.61?

Given the problems we're having with Autoconf 2.60 and M4 1.4.5,
perhaps it's time to issue Autoconf 2.61, with the NEWS item
prominently mentioning that 1.4.5 or later is now recommended.
Any objections?
Permalink | Reply | 2 comments |
headers
Paul Eggert | 7 Aug 2006 09:16
Favicon

Re: [GNU Autoconf 2.60] testsuite: 131 134 135 137 190 198 failed

Thanks for your bug report.

Pierre <pierre42d <at> 9online.fr> writes:

>> configure:1805: checking for Fortran compiler default output file name
>> configure:1832: g95    conftest.f  >&5
>> Warnings: 0  Errors: 0
>> configure:1835: $? = 0
>> configure:1881: result: b.out

That "b.out" is due to a bug in Autoconf; it incorrectly deduced the
name of the executable.  What is the name of the executable that your
GNU Fortran 95 compiler outputs?  That is, if you execute this shell
script:

        cat >conftest.f <<'EOF'
              program main

              end
        EOF
        g95 conftest.f

what file does your g95 compiler create?

I installed the following patch to fix the other bugs noted in your
report.  Thanks again.

2006-08-06  Paul Eggert  <eggert <at> cs.ucla.edu>

	Fix test suite failures reported by Pierre in
(Continue reading)

Permalink | Reply | 0 comments |
headers
Ralf Wildenhues | 7 Aug 2006 13:11
Picon
Picon

Re: any objections to Autoconf 2.61?

Hello Paul,

* Paul Eggert wrote on Mon, Aug 07, 2006 at 07:43:05AM CEST:
> Given the problems we're having with Autoconf 2.60 and M4 1.4.5,
> perhaps it's time to issue Autoconf 2.61, with the NEWS item
> prominently mentioning that 1.4.5 or later is now recommended.

Is "the problems" the handful of bug reports stemming from changed M4
warning output that causes a spurious test suite failure?  If so, I
think the test suite should pass with both 1.4.5 and previous M4
versions, even if they aren't recommended.  And I don't think it's
such a big issue, if you ask me (but maybe I have missed something
here).

Other than that, I'd like to do a test round before a release, and take
a look at the remaining changes; esp. the X rpath one looks a bit
worrying, at a glance, not having had a chance to look any closer yet.

But it's not me to decide, of course.

Cheers,
Ralf
Permalink | Reply | 1 comment |

Gmane