Horton

Re: Improving psql's \password command

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: "Bossart, Nathan" <bossartn@amazon.com>
Cc: "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Date: 2021-10-30T00:06:58Z
Lists: pgsql-hackers 
"Bossart, Nathan" <bossartn@amazon.com> writes:
> On 10/29/21, 12:47 PM, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>> While testing that, I noticed another bit of user-unfriendliness:
>> there's no obvious way to get out of it if you realize you are
>> setting the wrong user's password.  simple_prompt() ignores
>> control-C, and when you give up and press return, you'll just
>> get the prompt to enter the password again.

> Well, as of bf6b9e9, "ALTER ROLE nathan PASSWORD ''" is effectively
> the same as "ALTER ROLE nathan PASSWORD NULL".  I agree about the
> user-unfriendliness, but maybe simple_prompt() ignoring control-C is
> the root-cause of the user-unfriendliness.

I was afraid somebody would say that.  I have looked at it, but AFAICS
we'd have to duplicate all of sprompt.c and nearly all of pg_get_line.c
in order to tie it into psql's SIGINT infrastructure, since we wouldn't
dare enable the signal handler except during the innermost fgets() call,
and if we did get a signal we'd still need to clean up the terminal echo
state, so we couldn't just longjmp out of simple_prompt().  The
cost/benefit ratio of that doesn't look very good.

(Note that most callers of simple_prompt() don't need to sweat over
this because they haven't moved SIGINT handling off the default:
they're OK with just terminating on control-C.)

			regards, tom lane

Commits

  1. pg_receivewal, pg_recvlogical: allow canceling initial password prompt.

  2. Allow psql's other uses of simple_prompt() to be interrupted by ^C.

  3. Provide a variant of simple_prompt() that can be interrupted by ^C.

  4. Make psql's \password default to CURRENT_USER, not PQuser(conn).