Re: add non-option reordering to in-tree getopt_long

Michael Paquier <michael@paquier.xyz>

From: Michael Paquier <michael@paquier.xyz>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Kyotaro Horiguchi <horikyota.ntt@gmail.com>, noah@leadboat.com, pgsql-hackers@postgresql.org
Date: 2023-07-14T04:27:26Z
Lists: pgsql-hackers
On Thu, Jul 13, 2023 at 07:57:12AM -0700, Nathan Bossart wrote:
> Assuming you are referring to [0], it looks like you are missing 411b720.
> 
> [0] https://github.com/michaelpq/postgres/commits/getopt_test

Indeed, it looks like I've fat-fingered a rebase here.  I am able to
get a clean CI run when running this patch, sorry for the noise.

Anyway, this introduces a surprising behavior when specifying too many
subcommands.  On HEAD:
$ pg_ctl stop -D $PGDATA kill -t 20 start
pg_ctl: too many command-line arguments (first is "stop")
Try "pg_ctl --help" for more information.
$ pg_ctl stop -D $PGDATA -t 20 start
pg_ctl: too many command-line arguments (first is "stop")
Try "pg_ctl --help" for more information.

With the patch:
$ pg_ctl stop -D $PGDATA -t 20 start
pg_ctl: too many command-line arguments (first is "start")
Try "pg_ctl --help" for more information.
$ pg_ctl stop -D $PGDATA kill -t 20 start
pg_ctl: too many command-line arguments (first is "kill")
Try "pg_ctl --help" for more information.

So the error message reported is incorrect now, referring to an
incorrect first subcommand.
--
Michael

Commits

  1. Simplify option handling in pg_ctl.

  2. Teach in-tree getopt_long() to move non-options to the end of argv.