autoconf-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: four patches


From: Stepan Kasal
Subject: Re: four patches
Date: Thu, 30 Jun 2005 15:31:46 +0200
User-agent: Mutt/1.4.1i

Hello,

On Wed, Jun 29, 2005 at 05:00:51AM -0700, Noah Misch wrote:
> What is your rationale for these changes?

> > autoconf-20050629-subst-files.patch

Well, this was a silly mistake, sorry.

I have never used AC_SUBST_FILE, and I somehow thought that @foo@ is
replaced by file ./foo, so I thought $foo is not involved.

Only one isuue remains: the title "Output files" is unfortunate here.
I'd say "output files" are those declared with AC_CONFIG_FILES.

What about "File substitutions"?  (Patch attached.)

> > autoconf-20050629-precious.patch  changes the way precious variables are
> >     handled.  (Though its function doesn't depend on the previous patch,
> >     it won't apply smoothly without it.)

Currently, Autoconf generates four lines to store each precious variable.
And then it uses
        for `set | grep ^ac_env`
to verify all variables.

I thought it might be simpler to have a list of all precious variables, and
use the list both to store them and to check them.

I hope this explains most of the patch.[1]

And the code is not expanded if there are no precious variables.
"Why ?", you might ask.  "AC_INIT itself declares three precious variables."

Well, the situation is the same as with AC_SUBSTs.  Even though AC_INIT
declares several of them, the code is still ready for the situation when
there are none.

Looking forward to more comments,
        Stepan

[1] Why I said "most of the patch"?  Well, the patch contains one more
change: some code was moved from _AC_INIT_PREPARE to AC_INIT.
(When I realized that the patch actually contains two independent parts,
I was lazy to split the patch...)




reply via email to

[Prev in Thread] Current Thread [Next in Thread]