Re: jsonpath

On Thu, Mar 28, 2019 at 3:25 PM Andrew Dunstan
<andrew(dot)dunstan(at)2ndquadrant(dot)com> wrote:
> On 3/28/19 5:38 AM, Alexander Korotkov wrote:
> > On Thu, Mar 28, 2019 at 5:55 AM Andrew Dunstan
> > <andrew(dot)dunstan(at)2ndquadrant(dot)com> wrote:
> >> On 3/27/19 9:48 AM, Tom Lane wrote:
> >>> Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru> writes:
> >>>> Still no reproduction.
> >>> Annoying, but it's probably not worth expending more effort on
> >>> right now. I wonder whether that buildfarm animal can be upgraded
> >>> to capture core dump stack traces --- if so, then if it happens
> >>> again we'd have more info.
> >> I was able to get this stack trace.
> > Thank you very much! It appears to be hard to investigate this even
> > with backtrace. You told you can almost reliably reproduce this.
> > Could you please, find exact commit caused this error using "git
> > bisect"? I would very appreciate this.
>
> I'll try. It's time consuming given how long builds take.

Thank you very much for your efforts!

> Here's an interesting data point. If I run the whole jsonpath.sql script
> it crashes every time. If I run just the offending statement it crashes
> exactly every other time. It looks like in that case something gets
> clobbered and then cleared.

Could you clarify this a bit? What is exactly every other time? Do
you mean it doesn't crash first time, but crashes second time? Do you
run each offending statement in the separate session? Or do you run
multiple offending statements in the same session?

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company