perl post-wheezy

Dominic Hargreaves dom at earth.li
Thu May 9 11:07:31 UTC 2013


On Thu, May 09, 2013 at 01:52:48PM +0300, Niko Tyni wrote:
> Moving this public, we've been discussing things in private
> mostly accidentally. 
> 
> Summary: Perl 5.18-RC0 is out, so it looks like we're going to skip 5.16
> in sid/jessie. There's a debian-5.17 branch based on 5.17.11 on the git
> repo that will probably be renamed debian-5.18 soon.

Just to clarify, it's not quite out yet, but the final preparations
are being made.

> On Thu, May 09, 2013 at 10:39:42AM +0100, Dominic Hargreaves wrote:
> > On Thu, May 09, 2013 at 10:25:14AM +0300, Niko Tyni wrote:
> > > On Tue, May 07, 2013 at 11:35:56PM +0100, Dominic Hargreaves wrote:
> > > 
> > > > Okay, done. It needs much more testing (I haven't actually got further
> > > > than investigating the -Werror=format-security issue[1] yet) but I have
> > > > updated the dual-lived module deps. Copyright review still todo. Would
> > > > welcome a fresh set of eyes on the overall merging. My notes are attached
> > > > in case they are of use.
> 
> > > I had a look an updated debian/copyright (looking for 'copyright'
> > > and 'license' in the diff against 5.16, so I may have missed something.)
> > > Please have a look at the hv_func.h license mess and let me know if you
> > > think I'm overdoing it.
> > 
> > Will do.
> > 
> > > No testing yet.
> > > 
> > > The format-security thing cropping up on only i386 is weird.  I think I
> > > noticed something similar with the XS modules when the hardening flags
> > > were added. gcc clearly does something arch specific here. Possibly this
> > > is a bug worth reporting, particularly if there's an easy test case.
> > 
> > It's reported and there's a fix pending at
> > 
> > https://rt.perl.org/rt3/Ticket/Display.html?id=117893
> > 
> > Sorry that I missed that off my previous message!
> 
> No, I meant a report against gcc about the arch specific difference
> with format string warnings.

It's more complicated than that. gcc wants one thing, and g++ wants
the opposite!

https://rt.perl.org/rt3/Ticket/Display.html?id=117893#txn-1212949

> > > Some observations:
> > > 
> > > - debian/released-versions needs an update. The current version needs
> > >   to be there for it to appear in the perlapi-* substvar.
> > 
> > Ah yes, of course.
> > 
> > > - debian/patches/debian/skip-kfreebsd-crash.diff seems obsolete?
> > 
> > Okay, will check.
> > 
> > > - there are quite a few new deprecated modules. I found
> > >     Archive::Extract
> > >     B::Lint
> > >     CPANPLUS
> > >     Devel::InnerPackage
> > >     Log::Message
> > >     Module::Pluggable
> > >     Object::Accessor
> > >     Term::UI
> > >     Term::UI::History
> > >   These probably need to be separate packaged etc.
> > > - encoding.pm (in cpan/Encode/encoding.pm) is now deprecated;
> > >   it would probably be useful to grep for
> > >     Use of the encoding pragma is deprecated
> > >   in the build logs and file bugs once we're that far.
> > 
> > Yep. I think we have a reasonably well oiled machine for a lot of this
> > now, which is nice. There's a separate bug for deprecations already:
> > 
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=702096
> 
> Ah, yes. Thanks, I'll update that bug.
> 
> > > > http://perl5.git.perl.org/perl.git/commit/4f9ffe5913849a7e1b1b2efa4cce3e0d0ec3a790
> > > 
> > > Yay, so now we just rename the branch to debian-5.18 ? :)
> > 
> > Yep, I wasn't expecting debian-5.17 to stay around for a long time :)
> > 
> > I will pick up some next steps at the weekend, hopefully.
> 
> OK, I'll see what I can do too.

-- 
Dominic Hargreaves | http://www.larted.org.uk/~dom/
PGP key 5178E2A5 from the.earth.li (keyserver,web,email)




More information about the Perl-maintainers mailing list