Tom Lane writes:
> You seem to have missed the point completely - Neil wanted to suppress
> the normal progress messages *on stderr* so that he'd only get an email
> report when there was an abnormal event. That request seemed fine to
> me, though I thought the patch itself might have some portability
> issues. If it were recoded to follow the style of the existing boolean
> options in pg_dumpall then I'd recommend accepting it.
Two points:
1) The purpose of the progress messages is to inform the user of what
database he is connecting to. Previously, the user just got a prompt
"Password:" and had to guess which one. Unfortunately, it's hard for
pg_dumpall to predict whether a password prompt will appear.
This will become a nonissue soon because:
2) I'm currently rewriting pg_dumpall in C (as part of pg_dump actually),
so this patch will become obsolete.
(Also, I think we don't support external password files anymore, so the
password is now predictable and the progress messages can be removed, if
so.)
--
Peter Eisentraut peter_e@gmx.net
