Re: plan cache overhead on plpgsql expression

čt 19. 3. 2020 v 10:47 odesílatel Amit Langote <amitlangote09(at)gmail(dot)com>
napsal:

> Hi Pavel,
>
> Sorry it took me a while to look at this.
>
> On Tue, Feb 25, 2020 at 4:28 AM Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
> wrote:
> > po 24. 2. 2020 v 18:56 odesílatel Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
> napsal:
> >> But I found one issue - I don't know if this issue is related to your
> patch or plpgsql_check.
> >>
> >> plpgsql_check try to clean after it was executed - it cleans all plans.
> But some pointers on simple expressions are broken after catched exceptions.
> >>
> >> expr->plan = 0x80. Is interesting, so other fields of this expressions
> are correct.
> >
> > I am not sure, but after patching the SPI_prepare_params the current
> memory context is some short memory context.
> >
> > Can SPI_prepare_params change current memory context? It did before. But
> after patching different memory context is active.
>
> I haven't been able to see the behavior you reported. Could you let
> me know what unexpected memory context you see in the problematic
> case?
>

There was a problem with plpgsql_check after I applied this patch. It
crashed differently on own regress tests.

But I cannot to reproduce this issue now. Probably there was more issues
than one on my build environment.

So my questions and notes about a change of MemoryContext after patching
are messy. Sorry for noise.

Regards

Pavel

>
> --
> Thank you,
> Amit
>