next release (was: warning in AC_CHECK_DECL)

[Ralf Wildenhues]

Hi Eric,

[ limiting to autoconf ]

* Eric Blake wrote on Thu, Feb 16, 2006 at 09:54:27PM CET:
> 
> Either we need to update the various gnulib macros borrowed from CVS
> autoconf to override bugs in autoconf 2.59, or we need to release
> autoconf 2.60.

(Don't do the former, please, if you ask me.  No end of trouble, and
gnulib maintainers seem not very eager to fix macro discrepancies.)

> What are the remaining issues in the way of releasing autoconf 2.60?

I can only speak for myself, so here's a list of stuff that IMVHO would
be nice to have:

- I have a patch cooking to make AS_IF n-ary (hopefully done tonight)

- The m4_for* patch needs at least one more round of bugfixing
  (hint: try bounds that are arithmetic expressions themselves..)

- It would be good to document
    m4_split
    m4_normalize
    m4_defun/m4_require  (maybe not (yet))
  There are plenty other useful m4_* macros that are not documented yet.
  But if documenting them also means having to fix (many of) them, it
  would be best to leave that to after 2.60.  I can produce a list if
  desired.

- Fix detection of Fortran libs for Intel Fortran 7.1
  http://thread.gmane.org/gmane.comp.sysutils.autoconf.patches/2294
  I can probably look into this one; IIRC some newer compiler version
  did something slightly different though and needed taken care of as
  well.

- Easily recreating the last configure args:
  http://thread.gmane.org/gmane.comp.sysutils.autoconf.general/5719
  http://article.gmane.org/gmane.comp.sysutils.autoconf.bugs/4734
  Even a more limited solution than the ones proposed in above thread
  would already help.

I plan to work on above issues in the given order, but would of course
welcome anyone to beat me to it.  ;-)

Oh, IIRC it would need someone to actually produce (an alpha release
and) a release (no raised hand here).

Cheers,
Ralf