[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Perl 5.12 update

Hi -perl and -release,

I've been working on things related to Perl 5.12.0, and I thought it's
about time for an update.

The versions in experimental have settled on a stable ABI now. I ended
up enabling 64-bit integers on all architectures, but I don't expect
that to affect many packages in practice.

We don't seem to have armel and hppa experimental buildds running, but
I've successfully built 5.12.0~rc3-1/hppa and 5.12.0-2/armel on the porter
machines.  (The mips 5.12.0-2 failure is timing dependent and I'm rather
surprised it got triggered. A give-back would most probably succeed.)

Upstream is releasing 5.12.1 in the next few days. It's a bugfix release
with minor changes. I intend to upload it into experimental quickly when
it's released.

I expect 5.10->5.12 to be a smaller change than 5.8->5.10, and my rebuild
tests seem to support this. The bugs relevant to this transition that
I'm aware of can be seen at


which currently shows about 20 FTBFS bugs with 5.12.0, many of them

(I have rebuilt ~2150 packages against different configurations and
 release candidates of 5.12.0 on amd64 and x86; broadly those with names
 matching /perl/ or linking against libperl. I'll redo this with 5.12.1
 when it's ready.)

Archive-wide, worst blockers for the transition that I know of are
 - #581268 / #578547 : libdata-alias-perl breaks and has a big number of
   reverse dependencies via libdata-visitor-perl
 - #578628 / #580328 : libregexp-copy-perl breaks and makes much of the
   Catalyst suite uninstallable via libhtml-formfu-perl
 - #578814 : perl-tk breaks and has a big number of reverse dependencies
   (already fixed upstream in development releases)

(BTW, I don't have a good tool to list all recursive reverse dependencies
 and build dependencies. Any suggestions?)

Given all the above, I think it would be possible to get 5.12.1 in Squeeze
without severe delay effects on the release, as long as the worst blocker
issues are resolved somehow. However, staying at 5.10.1 and leaving 5.12
or 5.14 for the next release is also fine by me. (Upstream now plans to
do yearly major releases.)

Upstream support is probably going to be better for 5.12.1, but I doubt
there's much difference with security related issues. Ubuntu just released
their long time support version with 5.10.1, so we'd be in the same boat
for better or worse. OTOH, the Fedora people seem to be working hard to
get 5.12 in.

One complication with going for 5.12 in Squeeze is that I intend to be
unavailable for much of the summer. While I'm sure there are plenty of
people around who can patch the perl package in times of dire need, it
would probably be good to have somebody nominate themselves as a backup.

(My schedule is not set in stone yet, but let's discuss that in private
 if necessary.)

Release team, please let me know what you think.
Niko Tyni   ntyni@debian.org

Attachment: signature.asc
Description: Digital signature

Reply to: