Bug#758282: postgresql-9.3: FTBFS on sparc: 1 of 136 tests failed

[Christoph Berg]

Re: Niko Tyni 2014-08-16 <20140816092217.GA2953 at estella.local.invalid>
> Package: postgresql-9.3
> Version: 9.3.4-2
> Severity: serious
> 
> Rebuilding this package against Perl 5.20 failed on the sparc buildd,
> apparently due to a regression test failure.
> 
>  https://buildd.debian.org/status/fetch.php?pkg=postgresql-9.3&arch=sparc&ver=9.3.4-2%2Bb1&stamp=1408173187

I believe this is gcc-4.9's fault, also seen in postgresql-9.4. The
issue seems to be something like spilling constants and other
weirdnesses that only happen under register pressure. s390 has similar
issues in the "isolation" tests. The package builds/tests fine with
-O1. If the problems persist, I'll disable -O2 in postgresql-9.4's
debian/rules, but I'd prefer the compiler getting fixed :)

Re: Debian Bug Tracking System 2014-08-16 <handler.s.C.140818156531733.transcript at bugs.debian.org>
> Processing commands for control at bugs.debian.org:
> 
> > block 753529 with 758282 758283
> Bug #753529 [release.debian.org] transition: perl 5.20
> 753529 was blocked by: 758127 757571 757754 758270
> 753529 was not blocking any bugs.
> Added blocking bug(s) of 753529: 758283 and 758282

postgresql-9.3 needs to be removed from testing anyway as we are
moving to postgresql-9.4. If it helps with the perl transition, you
can remove it from testing now or as soon as the rdeps have
transitioned to 9.4. The current status is:

pgsql-asn1oid: RC bug with patch filed, waiting for maintainer
pg-reorg: Needs porting upstream, could just be removed from testing
  if it remains the only blocker
pgrouting: Waiting for maintainer, though I've heard he's working on
  it, so should happen soonish
postgis: Uploaded to sid, waiting for testing transition
(postgresql-pljava: is sid-only atm anyway)

I don't intend to do any further postgresql-9.3 uploads to sid.

Christoph
-- 
cb at df7cb.de | http://www.df7cb.de/