Re: [HACKERS] Walsender timeouts and large transactions

On 15 December 2017 at 00:36, Andrew Dunstan <andrew(dot)dunstan(at)2ndquadrant(dot)com
> wrote:

>
>
> On 12/14/2017 01:46 AM, Craig Ringer wrote:
> > On 7 December 2017 at 01:22, Petr Jelinek
> > <petr(dot)jelinek(at)2ndquadrant(dot)com <mailto:petr(dot)jelinek(at)2ndquadrant(dot)com>>
> > wrote:
> >
> > On 05/12/17 21:07, Robert Haas wrote:
> > >
> > > Generally we write if (a && b) { ... } not if (a) { if (b) .. }
> > >
> >
> > It's rather ugly with && because one of the conditions is two
> > line, but
> > okay here you go. I am keeping the brackets even if normally don't
> for
> > one-liners because it's completely unreadable without them IMHO.
> >
> >
> >
> > Yeah, that's why I passed on that FWIW. Sometimes breaking up a
> > condition is nice. Personally I intensely dislike the convention of
> >
> >
> > if (big_condition
> > && big_condition)
> > one_linerdo_something;
> >
> >
> > as awfully unreadable, but I guess code convention means you live with
> > things you don't like.
> >
> >
> > Anyway, I've just hit this bug in the wild for the umpteenth time this
> > year, and I'd like to know what I can do to help progress it to
> > commit+backport.
> >
> >
>
>
> Ask and ye shall receive. I've just committed it.
>
>
Brilliant, thanks. Backpatched too, great.

Now I'm going to shamelessly point you at the other nasty recurring logical
decoding bug while you're fresh from thinking about replication. I can
hope, right ;)

https://commitfest.postgresql.org/16/1397/ causes errors or bad data to be
decoded, so it's a serious bug.

I'll see if I can rustle up some review attention first though.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services