Horton

Re: pg11+: pg_ls_*dir LIMIT 1: temporary files .. not closed at end-of-transaction

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Justin Pryzby <pryzby@telsasoft.com>, pgsql-hackers@postgresql.org, Fabien COELHO <coelho@cri.ensmp.fr>, David Steele <david@pgmasters.net>, "Bossart, Nathan" <bossartn@amazon.com>, Thomas Munro <thomas.munro@gmail.com>
Date: 2020-03-12T14:05:38Z
Lists: pgsql-hackers 
Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> I wonder if this isn't saying that the whole value-per-call protocol is
> bogus, in that it seems impossible to write a useful function with it.

Only if you have a *very* narrow definition of "useful function".
If you look through SRF_RETURN_DONE callers, only a small minority
are trying to do resource cleanup beforehand.

> Maybe we should add one final call with a special flag "function
> shutdown" or something, so that these resources can be released if the
> SRF isn't run to completion?

We already have an appropriate mechanism for cleaning up resources,
ie RegisterExprContextCallback.  I do not think what you're suggesting
could be made to work without an incompatible break in the API for
SRFs, so it's not an improvement over telling people more forcefully
about how to do resource cleanup correctly.

			regards, tom lane

Commits

  1. Teach pg_ls_dir_files() to ignore ENOENT failures from stat().

  2. Remove useless pfree()s at the ends of various ValuePerCall SRFs.

  3. Avoid holding a directory FD open across assorted SRF calls.

  4. Document pg_ls_*dir hiding of directories and special files

  5. Avoid holding a directory FD open across pg_ls_dir_files() calls.

  6. Adjust nodeFunctionscan.c to reset transient memory context between calls