Re: Forgive trailing semicolons inside of config files

Isaac Morland <isaac(dot)morland(at)gmail(dot)com> writes:
> On Tue, 11 Jul 2023 at 10:43, Greg Sabino Mullane <htamfids(at)gmail(dot)com>
>> # Add settings for extensions here
>> random_page_cost = 2.5;
>>
>> Boom! Server will not start. Surely, we can be a little more liberal in
>> what we accept? Attached patch allows a single trailing semicolon to be
>> silently discarded.

> Please, no!

I agree. Allowing this would create huge confusion about whether it's
EOL or semicolon that ends a config file entry. If you can write a
semicolon, then why not spread an entry across lines, or write
multiple entries on one line?

It seems possible that someday we might want to convert over to
semicolon-is-end-of-entry precisely to allow such cases. But
I think that if/when we do that, it should be a flag day where you
*must* change to the new syntax. (We did exactly that in pgbench
scripts some years ago, and people didn't complain too much.)

> Or maybe there could be a "check configuration" subcommand which checks the
> configuration.

We have such a thing, see the pg_file_settings view.

regards, tom lane