[visit-developers] Funky changes to configure between revs 7178 and 7183

Mark Miller miller86 at llnl.gov
Wed May 20 12:54:27 EDT 2009


Hi Eric,

Yes, this is possible as 'autoreconf' uses m4 and aclocal each of which
MIGHT be installed separately from autoconf itself. For example, I
changed my path to include 2008 version of m4 and configure changed
MASSIVELY. I would expect that as m4 is doing most of the work under
autoconf.

Nonetheless, in this case, it is clear from looking at the code in
ac/with3rd.m4.

After talking with Hank, we think this may be due to use of 'autoconf'
instead of 'autoreconf'

Mark

On Wed, 2009-05-20 at 09:13 -0700, Eric Brugger wrote:
> Mark,
> 
> I noticed extra changes in the configure code when I changed
> configure.in and ran autoconf. The version numbers from autoconf
> matched, so I didn't understand why they were different. Perhaps
> the same version can produce different results on different
> machines (in other words, they aren't really the same). In any
> case, I took the path of least resistance and hand modified
> configure since my change was very straight forward.
> 
> Eric
> 
> At 09:01 AM 5/20/2009, Mark Miller wrote:
> >I neglected to mention that I am seeing similar issues with
> >visit-config.h.in. The location of symbols in that file has changed
> >solely as a result of my having run autoreconf.
> >
> >Mark
> >
> >On Wed, 2009-05-20 at 08:55 -0700, Mark Miller wrote:
> > > Hi all,
> > >
> > > I need help figuring out what I am seeing after having made a one line
> > > change to configure.in for dbio-only builds and re-generating configure
> > > using autoconf 2.59.
> > >
> > > I am seeing differences in the code that manages with --with-hdf5, --
> > > with-silo and --with-pdb switches. These are all controlled via the
> > > ac/with3rd.m4 macro in the 'ac' dir.
> > >
> > > The configure code I am generating is different from what is committed
> > > on the trunk. I am coming the conclusion that somehow, the configure
> > > code that is on the trunk was generated from some 'older' version of the
> > > ac/with3rd.m4 macro.
> > >
> > > I don't know how this could have happened and an trying to investigate.
> > > Does anyone recall having to do something out of the ordinary when
> > > changing configure in the last couple of weeks?
> > >
> > > Mark
> > >
> >--
> >Mark C. Miller, Lawrence Livermore National Laboratory
> >email: mailto:miller86 at llnl.gov
> >(M/T/W) (925)-423-5901 (!!LLNL BUSINESS ONLY!!)
> >(Th/F)  (530)-753-8511 (!!LLNL BUSINESS ONLY!!)
> 
-- 
Mark C. Miller, Lawrence Livermore National Laboratory
email: mailto:miller86 at llnl.gov
(M/T/W) (925)-423-5901 (!!LLNL BUSINESS ONLY!!)
(Th/F)  (530)-753-8511 (!!LLNL BUSINESS ONLY!!)



More information about the visit-developers mailing list