Re: [HACKERS] New psql input mode problems

> We have a winner!
>
> The new gets_fromFile strips the trailing newline, so an empty line in a
> file really comes in as an empty line. I don't see any reason why the
> check was done as it was in the first place, so the correct line in
> mainloop.c should be:
>
> /* No more input. Time to quit, or \i done */
> if (line == NULL)
> {
>
> as suggested.

Good. Already done.

> Here's an idea: Why don't the regression tests use a single-user postgres
> backend? That way you have even more control over internals, you can run
> it on an uninstalled database, and you don't rely on the particularities
> of the output of some obscure front-end.

It is generally unsafe because they don't share locking with other
backends, and some table are shared between databases.

>
> But I'm not going to tailor psql around the regression tests. That's the
> wrong direction. Just run it once with the old one and then with the new
> one and put those results in the current tree as a temporary solution.

Yes, I am sure that will be done soon.

--
Bruce Momjian | http://www.op.net/~candle
maillist(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026