Re: Refactoring of connection with password prompt loop for frontends

On 2019-Sep-26, Michael Paquier wrote:

> On Wed, Sep 25, 2019 at 05:47:39PM -0300, Alvaro Herrera wrote:
> > This patch has been absolutely overlooked by reviewers. Euler, you're
> > registered as a reviewer for it. Will you submit a review soon? :-)
> >
> > It does not currently apply, so if we get a rebase, that'd be good too.
>
> Here you go. There were four conflicts in total caused by the switch
> to the central logging infra for vacuumlo/oid2name and the
> introduction of recovery_gen.c. No actual changes except the rebase.

Hmm, you have an XXX comment that appears to need addressing; and I'd
add an explanation about the looping behavior to the comment atop the
new function.

I see that vacuumlo and scripts/common retain their "have_password"
variables. That seems to be so that they can reuse one for other
databases. But how does that work with the new routine?

I notice you changed the historical error message

> - pg_log_error("connection to database \"%s\" failed", database);

to

> + pg_log_error("could not connect to database \"%s\": %s",

The change seems appropriate to me.

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services