Re: Forgive trailing semicolons inside of config files

On Tue, Jul 11, 2023 at 11:04 AM Isaac Morland <isaac(dot)morland(at)gmail(dot)com>
wrote:

> Please, no!
>
> There is no end to accepting sloppy syntax. What next, allow "SET
> random_page_cost = 2.5;" (with or without semicolon) in config files?
>

Well yes, there is an end. A single, trailing semicolon. Full stop. It's
not a slippery slope in which we end up asking the AI parser to interpret
our haikus to derive the actual value. The postgresql.conf file is not some
finicky YAML/JSON beast - we already support some looseness in quoting or
not quoting values, optional whitespace, etc. Think of the trailing
semicolon as whitespace, if you like. You can see from the patch that this
does not replace EOL/EOF.

> I'd be more interested in improvements in visibility of errors. For
> example, maybe if I try to start the server and there is a config file
> problem, I could somehow get a straightforward error message right in the
> terminal window complaining about the line of the configuration which is
> wrong.
>

That ship has long since sailed. We already send a detailed error message
with the line number, but in today's world of "service start", "systemctl
start", and higher level of control such as Patroni and Kubernetes, getting
things to show in a terminal window isn't happening. We can't work around
2>&1.

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

There are things inside of Postgres once it has started, but yeah,
something akin to visudo would be nice for editing config files.

> But I think it would be way more useful than a potentially never-ending
> series of patches to liberalize the config parser.
>

It's a single semicolon, not a sign of the parser apocalypse. I've no plans
for future enhancements, but if they do no harm and make Postgres more user
friendly, I will support them.

Cheers,
Greg