On 01/15/2012 05:33 PM, Jim Meyering wrote: > Stefano Lattarini wrote: > >> Hi Jim. >> >> On 01/15/2012 05:22 PM, Jim Meyering wrote: >>> Without this change, numerous tests would fail. >>> E.g., on a Fedora 16 system, running autoreconf would print this warning: >>> >>> Use of uninitialized value $ARGV[0] in pattern match (m//) at \ >>> /p/share/autoconf/Autom4te/General.pm line 273. >>> >> Ouch, apparently you have missed my earlier patch that fixed this same bug: >> <http://lists.gnu.org/archive/html/autoconf-patches/2012-01/msg00010.html> > > Hi Stefano, > > I did indeed miss that. Sorry. > > It didn't even occur to me that a known bug like that would > be left (awaiting review?) > Yes, awaiting review. For projects where I only contribute minimally, I want to wait for an explicit ACK before *any* commit, whether bug-fix or not. That's why I'm also waiting for a review before pushing this patch:
<http://lists.gnu.org/archive/html/autoconf-patches/2012-01/msg00011.html> which fix a bug more difficult to spot but which, when hit, can have much more heinous consequences (details in the commit log). > in master. > > I see you would have fixed it like this: > > if (@ARGV > 0 && $ARGV[0] =~ /^-./) > > I prefer that to what I did (defined $ARGV[0]), > but prefer this even more: > > if (@ARGV && $ARGV[0] =~ /^-./) > > It's concise (less syntax), and no use of ">". > I like this more as well. I assume you'll go ahead and make the change yourself, right? Regards, Stefano