[ Cc'ed to automake, see last point ]
Akim Demaille wrote:
> it seems that the primary issue is the fact that some macros need to be
> before AC_INIT, while most others should not.
correct.
> all (y)our troubles will vanish if *any* macro works after AC_INIT.
correct.
> I've always preferred this, included for AC_REVISION, kill me if you
> disagree.
Pfu! (Gun sound with a silencer). I stylistically prefer some of them
before AC_INIT, but style is definitely not the problem here.
> can PACKAGE and VERSION be computed at sh time. Personally, I'm extremely
> hostile to this
agreed.
> So my plan is to have AC_PACKAGE work after AC_INIT only, with a
> literal (in the sh sense: you still can m4-compute the arguments if
> you wish, forgive me if you don't).
That's fine with me.
> Alternatively, should AC_PACKAGE be removed in favor of AC_INIT/2-3
> args? I like this.
Yes! I even prefer this to calling AC_PACKAGE after AC_INIT, provided
that aclocal is loadable (and loaded) before AC_INIT.
While we're at it, the package name and version stuff should ideally
not be automake's AM_INIT_AUTOMAKE job. AM_INIT_AUTOMAKE's arguments could
then be made optional, and if not provided, it could use AC_PACKAGE_NAME and
AC_PACKAGE_VERSION.
--
/ / _ _ Didier Verna http://www.inf.enst.fr/~verna/
- / / - / / /_/ / EPITA / LRDE mailto:[EMAIL PROTECTED]
/_/ / /_/ / /__ / 14-16 rue Voltaire Tel. +33 (1) 44 08 01 77
94276 Kremlin-Bicêtre cedex Fax. +33 (1) 44 08 01 99
