Re: Fix chdir-long.m4 caching

[Ralf Wildenhues]

[ dropping bug-gnulib ]

Hello Eric,

* Eric Blake wrote on Tue, Sep 26, 2006 at 04:39:22AM CEST:
> According to Ralf Wildenhues on 9/25/2006 2:03 PM:
> > 
> >>       7 checking whether this system has an arbitrary file name length 
> >> limit... yes
> > 
> > OK to apply this patch?  I think this variable isn't used elsewhere in
> > gnulib nor coreutils, so renaming this shouldn't be a problem.
> 
> How about covering our tracks so this won't happen again?

Nice idea (I had it too but wasn't sure whether it could be done safely).
Your patch has the logic the other way round, and it fails to work with
shell indirections.  I'm still not sure whether the following is safe.

Cheers,
Ralf

Index: lib/autoconf/general.m4
===================================================================
RCS file: /cvsroot/autoconf/autoconf/lib/autoconf/general.m4,v
retrieving revision 1.930
diff -u -r1.930 general.m4
--- lib/autoconf/general.m4     14 Sep 2006 09:40:59 -0000      1.930
+++ lib/autoconf/general.m4     26 Sep 2006 04:58:17 -0000
@@ -1905,7 +1905,11 @@
 # The name of shell var CACHE-ID must contain `_cv_' in order to get saved.
 # Should be dnl'ed.  Try to catch common mistakes.
 m4_defun([AC_CACHE_VAL],
-[m4_bmatch([$2], [AC_DEFINE],
+[AS_LITERAL_IF([$1],
+  [m4_bmatch([$1], [_cv_], [],
+          [AC_DIAGNOSE([syntax],
+[$0($1, ...): suspicious cache-id, must contain _cv_ to be cached])])])dnl
+m4_bmatch([$2], [AC_DEFINE],
           [AC_DIAGNOSE([syntax],
 [$0($1, ...): suspicious presence of an AC_DEFINE in the second argument, ]dnl
 [where no actions should be taken])])dnl