Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc) - Mailing list pgsql-bugs

From Andres Freund
Subject Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc)
Date
Msg-id 20160413175827.dmlbtdf7c3mgmnex@alap3.anarazel.de
Whole thread Raw
In response to Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc)  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc)  (Christoph Berg <myon@debian.org>)
List pgsql-bugs
On 2016-04-13 13:54:10 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2016-04-13 11:41:17 +0200, Christoph Berg wrote:
> >> I've just uploaded 9.4.7 for Debian's wheezy-backports suite.
> >> Unfortunately it's now crashing twice in the contrib/test_decoding
> >> regression tests on sparc. I could reproduce the error on the porter
> >> box:
>
> > Hrmpf, I guess this is an alignment issue.
>
> Maybe, but then why didn't the buildfarm detect it?  We certainly
> have several critters that would be sticky about this.

I'm wondering the same. My theory is that it might only actually matter
when built optimized; which few animals do. The instruction selection
definitely can influence whether an alignment issue becomes aparent or
not.  Unless that idea pans out I guess I'll have to go through the
motions to get access to that porter box. I had that at some point in
the past (debugging something turning out to be a compiler bug).

Andres

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc)
Next
From: Christoph Berg
Date:
Subject: Re: Bus error in pg_logical_slot_get_changes (9.4.7, sparc)