Following is not a bug...yet. It may eventually become one if/when perl 5.10.0 moves from overlays into the tree. While I was emerging a package (it seem to merge fine, btw.), I noticed strange warnings: Constant subroutine main::LC_ALL redefined at /usr/lib/perl5/5.10.0/Exporter.pm line 66. at /usr/bin/help2man line 38 Prototype mismatch: sub main::LC_ALL () vs none at /usr/lib/perl5/5.10.0/Exporter.pm line 66. at /usr/bin/help2man line 38 When I googled for it, I came across following mail: http://article.gmane.org/gmane.comp.gnu.core-utils.bugs/13511 I don't know, if this breaks anything for <perl-5.10.0.
So a problem in sys-apps/help2man is fixed that would only arise when a version of dev-lang/perl hits the tree which isn't there yet? When is dev-lang/perl-5.10.0 due?
How exactly does this block perl-5.10? I don't see how this application blocks perl-5.10 from functioning properly; it just looks like an upstream bug unrelated to #206455. Read the patch message: Constant subroutine main::LC_ALL redefined at /.../Exporter.pm * man/help2man: Don't include LC_ALL in the "use POSIX" list, since Locale::gettext->import will get it. This shouldn't block anything.
I took a look at that mailing list again, shortly after the first patch got there, a corrected patch arrived. (it's http://news.gmane.org/find-root.php?message_id=%3c87d4o0e5om.fsf%40rho.meyering.net%3e). While those warning were non-fatal for the package I saw them in, I can't say if that will be the case for all packages. I don't know perl, but the fact that this patch was created suggests that it may sometimes be fatal.
(In reply to comment #2) > How exactly does this block perl-5.10? I don't see how this application blocks > perl-5.10 from functioning properly; it just looks like an upstream bug > unrelated to #206455. The stability of other packages is something to consider before a package they depend on goes into the tree. In this case sys-apps/help2man would probably need to be patched before dev-lang/perl-5.10 does come in, so one bug report quite naturally blocks the other.
(In reply to comment #4) > (In reply to comment #2) > > How exactly does this block perl-5.10? I don't see how this application blocks > > perl-5.10 from functioning properly; it just looks like an upstream bug > > unrelated to #206455. > > The stability of other packages is something to consider before a package they > depend on goes into the tree. In this case sys-apps/help2man would probably > need to be patched before dev-lang/perl-5.10 does come in, so one bug report > quite naturally blocks the other. > That would be valid except the patch posted looks like this bug was present with or without perl-5.10. My point is it was already a bug, but now it's much more evident with perl-5.10. Those POSIX symbols that the script is pulling are not new, and judging by the patch, did not change either. Since it's an upstream bug unrelated to perl-5.10 that existed before perl-5.10, it's not of my opinion that this is a blocker.
you should at least link to the correct patch instead of a broken one ... then again, neither actually apply has anyone here confirmed that the changes in question work ? --- help2man.PL +++ help2man.PL @@ -34,7 +34,7 @@ my $have_gettext; BEGIN { eval { require Locale::gettext; - Locale::gettext->import; + Locale::gettext->import (qw(gettext textdomain)); $have_gettext = 1; };
Debian made a change to Locale::gettext itself instead of every app which uses it: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=479803
that makes more sense to me
Fixed in 1.05-r1. Thanks all.