Re: Detoasting optionally to make Explain-Analyze less misleading

Matthias van de Meent <boekewurm+postgres@gmail.com>

From: Matthias van de Meent <boekewurm+postgres@gmail.com>
To: Tomas Vondra <tomas.vondra@enterprisedb.com>
Cc: stepan rutz <stepan.rutz@gmx.de>, Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@postgresql.org
Date: 2024-02-26T19:30:56Z
Lists: pgsql-hackers

Attachments

Hi,

I've taken the liberty to update this patch, and register it in the
commitfest app to not lose track of progress [0].

The attached v8 patch measures scratch memory allocations (with MEMORY
option), total time spent in serialization (with TIMING on, measures
are inclusive of unseparated memcpy to the message buffer), and a
count of produced bytes plus the output format used (text or binary).
It's a light rework of the earlier 0007 patch, I've reused tests and
some infrastructure, while the implementation details and comments
have been updated significantly.

I think we can bikeshed on format and names, but overall I think the
patch is in a very decent shape.

Stepan, thank you for your earlier work, and feel free to check it out
or pick it up again if you want to; else I'll try to get this done.

Kind regards,

Matthias van de Meent

[0] https://commitfest.postgresql.org/47/4852/

Commits

  1. Invent SERIALIZE option for EXPLAIN.