On Mon, Feb 06, 2023 at 04:23:02PM +0900, Michael Paquier wrote:
> On top of that, I have noticed an extra combination that would not
> make sense and that could be checked with the SQL queries:
> GUC_DISALLOW_IN_FILE implies GUC_NOT_IN_SAMPLE. The opposite may not
> be true, though, as some developer GUCs are marked as
> GUC_NOT_IN_SAMPLE but they are allowed in a file. The only exception
> to that currently is config_file. It is just a special case whose
> value is enforced at startup and it can be passed down as an option
> switch via the postgres binary, still it seems like it would be better
> to also mark it as GUC_NOT_IN_SAMPLE? This is done in 0002, only for
> HEAD, as that would be a new check.
0001 has been applied to clean up the existing situation. Remains
0002, that I am letting sleep to see if there's interest for it, or
perhaps more ideas around it.
--
Michael