Re: Server crash with older tzload library

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Jeevan Chalke <jeevan(dot)chalke(at)enterprisedb(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Server crash with older tzload library
Date: 2010-03-11 18:49:15
Message-ID: 26096.1268333355@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Jeevan Chalke <jeevan(dot)chalke(at)enterprisedb(dot)com> writes:
> The upstream zoneinfo project just released a new tzcode version, 2010c.
> After syncing the code to this version does not lead to server crash. The
> new release is now initializing the tzstate variable with zeros to avoid any
> garbage values.

> PFA, patch which will bring us up-to date to 2010c.

I've applied the update to 2010c since that apparently fixes some
misbehaviors in obscure time zones (where is America/Godthab???).
However, the proposed addition of explicit clears of the tzstate
struct doesn't match any upstream change that I can see. I inserted
explicit initializations to random data instead and still couldn't
provoke a crash. While it seems harmless enough to explicitly zero
it, I'd like to see an instance of the reported crash, because I have
a feeling that the real problem you're dealing with is elsewhere.
If you can't provoke it reliably, maybe the zeroing didn't really
fix it.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Jeff Davis 2010-03-11 19:00:59 Re: operator exclusion constraints
Previous Message Pavel Stehule 2010-03-11 18:32:53 tsearch - using a transformed data files