Bug#659075: [rt.cpan.org #61577] ->sockdomain and ->socktype undefined on newly ->accept'ed sockets

[Niko Tyni]

On Sun, Apr 08, 2012 at 04:22:57PM +0100, Dominic Hargreaves wrote:
> On Sat, Apr 07, 2012 at 11:51:08AM +0200, intrigeri wrote:
> > I confirm Daniel's v4 patch fixes Debian#607674 and RT#68282 for me,
> > and eventually allows me to run a recent Net::Server.
> > 
> > Thanks, Daniel, for your persistence in getting this bug fixed both
> > upstream and in Debian.

Indeed, thank you Daniel.

> Okay - it looks like the only outstanding issue that's been expressed
> on this bug or on the CPAN bug is about Darwin compatibility; this
> isn't a current issue for Debian. So I'm inclined to apply this patch
> assuming it doesn't break the test suite. Any other opinions (Niko?)

Our long-standing general rule about upstream divergences in the Debian
perl package is that any non-Debian specific changes should go in blead
or maint-* first.

However, I can see obvious timing problems in this particular case:
bleadperl is currently in a code freeze for the 5.16 release, and Debian
will soon(ish) enter one for the wheezy release.

So yes, I think the sensible choice is to break the first rule and
apply the patch for the Debian 5.14 package. I'm not very happy about
that though, and I'm slightly worried that the issue will just stall
afterwards (because Daniel no longer has an incentive for pushing it :)

I suppose it would help a bit if there was an rt.perl.org ticket about
this; AFAIK there's currently just [rt.cpan.org #61577], which is
in the wrong place as blead is upstream for IO.
-- 
Niko Tyni   ntyni at debian.org