From: | David Rowley <dgrowleyml(at)gmail(dot)com> |
---|---|
To: | Justin Pryzby <pryzby(at)telsasoft(dot)com> |
Cc: | PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Andres Freund <andres(at)anarazel(dot)de> |
Subject: | Re: jit and explain nontext |
Date: | 2020-10-15 01:02:15 |
Message-ID: | CAApHDvpPt=iN=f5gHQSCA073mp5KO46gsEZFgTWc4xbzKshb9g@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, 15 Oct 2020 at 08:39, Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
>
> /* don't print information if no JITing happened */
> if (!ji || ji->created_functions == 0)
> return;
>
> This applies even when (es->format != EXPLAIN_FORMAT_TEXT), which I think is
> wrong. Jit use can be determined by cost, so I think jit details should be
> shown in non-text format whenever ji!=NULL, even if it's zeros. Arguably, bits
> could be omitted if jit_expressions=off or jit_tuple_deforming=off, but I don't
> see the point.
Just for some reference. Some wisdom was shared in [1], which made a
lot of sense to me.
If we apply that, then we just need to decide if displaying any jit
related fields without any jitted expressions is relevant.
I'm a bit undecided.
[1] https://www.postgresql.org/message-id/2276865.1593102811%40sss.pgh.pa.us
From | Date | Subject | |
---|---|---|---|
Next Message | Andy Fan | 2020-10-15 01:04:24 | Re: ALTER TABLE .. DETACH PARTITION CONCURRENTLY |
Previous Message | Andres Freund | 2020-10-15 00:56:16 | Re: gs_group_1 crashing on 13beta2/s390x |