> On Nov 19, 2014, at 3:31 PM, Eric Blake <ebl...@redhat.com> wrote: > > On 11/18/2014 08:54 PM, KO Myung-Hun wrote: >> -Zargs-resp enables a response file support on OS/2. >> >> * configure.ac (LDFLAGS): add -Zargs-resp on *-*-os2*. >> --- >> configure.ac | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/configure.ac b/configure.ac >> index 6dca663..2ca1576 100644 >> --- a/configure.ac >> +++ b/configure.ac >> @@ -217,6 +217,8 @@ M4_WITH_DMALLOC >> >> AC_CONFIG_COMMANDS([stamp-h], [[test -z "$CONFIG_HEADERS" || date > >> stamp-h]]) >> >> +AS_CASE([$host], [*-*-os2*], [LDFLAGS="$LDFLAGS -Zargs-resp"]) >> + > > You cannot assume $host is set unless you also guarantee that > AC_CANONICAL_HOST was called. Gnulib happens to call it, but I wonder > if we should also explicitly call it prior to this line so that someone > copying it to another file doesn't run into problems.
I re-organised this patch slightly to avoid being overridden by the user's LDFLAGS setting... but I only used it for builds from the src directory, which might not be the intent of the original patch. I'm not sure it's worth worrying about problems caused by wanton cut-n-pasting from configure.ac without understanding all the implications, or even that a manual AC_CANONICAL_HOST is the solution -- after all, someone can still cut-n-paste the one without the other into another project. Eric, by all means, if you disagree go ahead and make any additional changes you consider sensible. Cheers, -- Gary V. Vaughan (gary AT gnu DOT org) _______________________________________________ M4-patches mailing list M4-patches@gnu.org https://lists.gnu.org/mailman/listinfo/m4-patches
