Re: Align memory context level numbering in pg_log_backend_memory_contexts()

From: David Rowley <dgrowleyml(at)gmail(dot)com>
To: torikoshia <torikoshia(at)oss(dot)nttdata(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Align memory context level numbering in pg_log_backend_memory_contexts()
Date: 2025-04-15 21:03:45
Message-ID: CAApHDvpE6JS9F5kg+W4+p9TsiqFsPE+6XqEtD0P0HKNBoa9tYA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, 16 Apr 2025 at 01:23, torikoshia <torikoshia(at)oss(dot)nttdata(dot)com> wrote:
> =# select name, level from pg_backend_memory_contexts;
> name | level
> ----------------------------+-------
> TopMemoryContext | 1

> =# select pg_log_backend_memory_contexts(pg_backend_pid());
>
> LOG: level: 0; TopMemoryContext: ...

> However, for the sake of consistency and to avoid potential confusion,
> would it make sense to also start the levels from 1 in
> pg_log_backend_memory_contexts() starting in v18?

That's not a very nice inconsistency.

As for which one is changed... Quite a bit of thought and discussion
occurred before 32d3ed81 to try to make the "path" and "level" columns
as easy to use as possible, and making "level" 1-based was done as a
result of trying to make queries that sum up a context and its
children as easy as possible. The example at the end of [1] would be
a little more complex with 0-based levels as we'd have to use level+1
as the index.

My vote is to make the levels 1-based in all locations where we output
the context information.

Does anyone else have views on this?

David

[1] https://www.postgresql.org/docs/devel/view-pg-backend-memory-contexts.html

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Alvaro Herrera 2025-04-15 21:10:26 Re: not null constraints, again
Previous Message David Rowley 2025-04-15 20:49:46 Re: Add estimated hit ratio to Memoize in EXPLAIN to explain cost adjustment