Re: CVS trunk testing results

[Mark D. Baushke]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Derek Robert Price <derek@ximbiot.com> writes:

> Mark D. Baushke wrote:
> 
> > Larry Jones <lawrence.jones@eds.com> writes:
> >
> > Mark D. Baushke writes:
> >
> > >I wrote replacement functions for both ftello and fseeko in the lib
> > >directory. I ended up adding ftello and fseeko into an AC_REPLACE_FUNCS
> > >instead of using the AC_FUNC_FSEEKO macro when I couldn't get the latter
> > >to do the right thing in the lib/Makefile.in file.
> >
> > That's the problem.  On HP-UX, you have to #define _LARGEFILE_SOURCE to
> > make the fseeko/ftello prototypes visible and to correctly map them to
> > either the 32 or 64 bit library functions.  AC_FUNC_FSEEKO does that,
> > AC_REPLACE_FUNCS does not.
> >
> >
> > Hmmm... AC_FUNC_FSEEKO seems to expect folks to have a .c or .h file
> > that will conditionally define the fseeko and/or ftello functions rather
> > than adding a replacement .o file to a library.
> >
> > Should I just make them be a static functions that I put into a .h file
> > using #if !defined() or is there another way to deal with this?
> 
> 
> What about a simple:
> 
> 
> AC_FUNC_FSEEKO
> if test $ac_cv_func_fseeko = no; then
>     AC_LIBOBJ(fseeko);
>     AC_LIBOBJ(ftello);
> fi
> 
> 
> ? You would still need to #ifdef the prototypes.

Yes, that might work. The problem is that I don't have a system
that will exercise that part of the configure script, so I am
shooting in the dark...

        -- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQE/13ip3x41pRYZE/gRAijLAJ4sfS9lweE8ryzZWLo7w4g5OnQ7BQCeIDTA
G7gbS4AWZRwmLNP/58BK9Vo=
=ggie
-----END PGP SIGNATURE-----