On Tue, Jun 20, 2006 at 08:03:23AM +0200, demerphq wrote: > On 6/20/06, Ron Savage <[EMAIL PROTECTED]> wrote: > >On Mon, 19 Jun 2006 14:32:10 +0200, demerphq wrote: > > > >Hi demerphq > > > >> Nope. If you upgrade it should go in to site/lib, meaning you have > >> to use UNINST=1 if you upgrade. > > > >Are you sure it'll go into site/lib? > > Yep. Thats what the "shadowing" support in ExtUtils::Install is for, > and why UNINST=1 exists. (There would be no need for UNINST if new > versions predictaby overwrote the old ones.) > > >> At least thats what happens with things like Data::Dumper > > > >The module I love to hate. Some other modules eval its version number as a > >numeric, and report many errors because 2.121_04 is not numeric. Death to > >dev > >versions! > > Yeah, thats not all that unusual unfortunately. Maybe somebody should > put together a patch for blead so $VERSION is special cased not to > produce this warning.
The warning is proper, as it alerts you that perl isn't treating the "number" as a number. It would be a bad thing to just suppress it: $ cat Foo.pm package Foo; our $VERSION = '1.23_04'; 1; $ perl -we'use Foo 1.23_05' Argument "1.23_04" isn't numeric in subroutine entry at -e line 1. Foo version 1.2305 required--this is only version 1.23_04 at -e line 1. BEGIN failed--compilation aborted at -e line 1. $ perl -we'use Foo 1.23_04' Argument "1.23_04" isn't numeric in subroutine entry at -e line 1. Foo version 1.2304 required--this is only version 1.23_04 at -e line 1. BEGIN failed--compilation aborted at -e line 1. $ perl -we'use Foo 1.23_03' Argument "1.23_04" isn't numeric in subroutine entry at -e line 1. Foo version 1.2303 required--this is only version 1.23_04 at -e line 1. BEGIN failed--compilation aborted at -e line 1. $ perl -we'use Foo 1.23' Argument "1.23_04" isn't numeric in subroutine entry at -e line 1. $
