bug-autoconf
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CVS autoconf testsuite failure on NetBSD/amd64


From: Nicolas Joly
Subject: Re: CVS autoconf testsuite failure on NetBSD/amd64
Date: Wed, 15 Dec 2004 17:09:59 +0100
User-agent: Mutt/1.5.6i

On Wed, Dec 15, 2004 at 04:20:53PM +0100, Stepan Kasal wrote:
> Hi,
> 
> On Wed, Dec 15, 2004 at 02:17:49PM +0100, Nicolas Joly wrote:
> > The second one is new. I tracked this one a little and noticed a small
> > problem in `_AC_FEATURE_CHECK_LENGTH'.
> ...
> > -  done])
> > +  done
> > +  rm -f conftest.*])
> 
> I introduced this change.  I didn't add the rm because all conftest*
> files should be deleted on exit by the ``trap'' defined in general.m4
> in _AC_INIT_PREPARE .  Why did this fail to clean the files?

The files are indeed cleaned at the end of the configure script.

address@hidden [testsuite.dir/156]> ./configure
[...]
+ test 0 != 0
+ echo configure: exit 0
+ rm -f core *.core
+ rm -f -r conftest.in conftest.nl conftest.out confdefs.h conf23761*
+ exit 0
address@hidden [testsuite.dir/156]> 

But the autotest framework add some extra hooks (AC_STATE_SAVE) in the
configure.ac:

address@hidden [testsuite.dir/156]> cat configure.ac 
AC_INIT
AC_CONFIG_AUX_DIR($top_srcdir/config)
AC_CONFIG_HEADER(config.h:config.hin)
AC_PROG_GREP
AC_STATE_SAVE(before)
AC_PROG_SED
AC_OUTPUT
AC_STATE_SAVE(after)

This way it checks that autoconf macros do not forget to clean
temporary files they create.

-- 
Nicolas Joly

Biological Software and Databanks.
Institut Pasteur, Paris.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]