Horton

Re: Fix some error handling for read() and errno

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Postgres hackers <pgsql-hackers@postgresql.org>, Robert Haas <robertmhaas@gmail.com>, Magnus Hagander <magnus@hagander.net>, Robbie Harwood <rharwood@redhat.com>, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>, Tom Lane <tgl@sss.pgh.pa.us>, hlinnaka <hlinnaka@iki.fi>
Date: 2018-07-16T11:09:19Z
Lists: pgsql-hackers 
On Sat, Jul 14, 2018 at 03:37:56PM +0900, Michael Paquier wrote:
> For now, I think that just moving forward with 0001, and then revisit
> 0002 once the other 2PC patch is settled makes the most sense.  On the
> other thread, the current 2PC behavior can create silent data loss so
> I would like to back-patch it, so that would be less work.

Are there any objections with this plan?  If none, then I would like to
move on with 0001 as there is clearly a consensus to simplify the work
of translators and to clean up the error code paths for read() calls.
Let's sort of the rest after the 2PC code paths are addressed.
--
Michael

Commits

  1. Rework error messages around file handling