Re: cygwin libtool breakage

[Gary V . Vaughan]

On Tuesday 05 June 2001  1:04 am, Robert Collins wrote:
> I added AC_REQUIRE([AC_PROG_CC])dnl to
> /usr/share/autoconf/acspecific.m4, and it made no difference at all. I
> did several variations-on-a-theme, with no luck.
>
> Adding
> AC_EXEEXT immediately _after_ the AC_PROG_CC line in configure.in for
> libtool, made it behave.
>
> I've no idea _why_ this is so, perhaps some m4 guru would like to
> investigate? I'm happy to spend time on this, if guided, or ship data
> elsewhere.

The AC_REQUIRE machinery in 2.13 is a bit flaky at times.  

I may be way off base here since I haven't actually looked at the code, but I 
think that all of the AC_REQUIRED code is built in a diversion which is 
undiverted into the configure script before the macros that are expanded 
inline.  It looks as though the requirements graph has confused autoconf, and 
the AC_EXEEXT macro (and a few others too if you examine the current 
ordering) have been placed too early in the configure script.  By explicitly 
naming AC_EXEEXT, any AC_REQUIRES are automatically satisfied and the macro 
doesn't arrive in the diversion.

It seems to me that the correct fix (wrt autoconf-2.13) is indeed to put an 
explicit call in configure.in.  If I am along the right lines, it seems 
likely that 2.50 won't exhibit this problem.  Unless we find out something to 
the contrary, I'll apply your fix right after I'm done with the merge.

Cheers,
        Gary.
-- 
  ())_.  Gary V. Vaughan     gary@(oranda.demon.co.uk|gnu.org)
  ( '/   Research Scientist  http://www.oranda.demon.co.uk        ,_())____
  / )=   GNU Hacker          http://www.gnu.org/software/libtool   \'      `&
`(_~)_   Tech' Author        http://sources.redhat.com/autobook    =`---d__/