Horton

Re: AlterSubscription_refresh "wrconn" wrong variable?

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Peter Smith <smithpb2250@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Peter Eisentraut <peter_e@gmx.net>
Date: 2021-05-04T04:31:49Z
Lists: pgsql-hackers 
Hi,

On 2021-05-04 09:29:42 +1000, Peter Smith wrote:
> While reviewing some logical replication code I stumbled across a
> variable usage that looks suspicious to me.

> Note that the AlterSubscription_refresh function (unlike other
> functions in the subscriptioncmds.c) is using the global variable
> "wrconn" instead of a local stack variable of the same name. I was
> unable to think of any good reason why it would be deliberately doing
> this, so my guess is that it is simply an accidental mistake that has
> gone unnoticed because the compiler was silently equally happy just
> using the global var.

> Apparently, this is not causing any reported problems because it seems
> like the code has been this way for ~4 years [1].

This sounded vaguely familiar. After a bit of searching I found that's
because I debugged a crash related to it:
https://www.postgresql.org/message-id/20201111215820.qihhrz7fayu6myfi%40alap3.anarazel.de

Peter?

Greetings,

Andres Freund

Commits

  1. Rename the logical replication global "wrconn"

  2. AlterSubscription_refresh: avoid stomping on global variable

  3. Logical replication support for initial data copy