Re: IO wait events for COPY FROM/TO PROGRAM or file

Nikolay Samokhvalov <nik@postgres.ai>

From: Nikolay Samokhvalov <nik@postgres.ai>
To: Michael Paquier <michael@paquier.xyz>
Cc: Sami Imseih <samimseih@gmail.com>, pgsql-hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-01-31T19:51:39Z
Lists: pgsql-hackers

Attachments

On Sun, Jan 11, 2026 at 02:22 Michael Paquier <michael@paquier.xyz> wrote:

> On Thu, Jan 08, 2026 at 09:22:26PM -0600, Sami Imseih wrote:
> > I don't see any issue with this approach for fwrite().
>
> Please feel free to discard my comment, then :)
>
> > Another comment. Wouldn't be better to use "COPY FROM" and "COPY TO" in
> the
> > names to make it more obvious they are related to the COPY command?
>
> Yeah, we had better do that.  That's slightly cleaner for the user if
> they mix both COPY types at the same time, not requiring a mental
> mapping that FROM is a read and TO is a write.



v2 attached with the rename per feedback: COPY_DATA_READ/WRITE →
COPY_FROM_READ/COPY_TO_WRITE.

Given how small this patch is, any chance it could still make PG19?

>

Commits

  1. Add two IO wait events for COPY FROM/TO on a pipe/file/program