Re: TAP / recovery-test fs-level backups, psql enhancements etc

Craig Ringer wrote:
> On 3 March 2016 at 21:16, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote:

> > > The rest are feature patches:
> > > 004 allows filtering on RecursiveCopy by a predicate function. Needed for
> > > filesystem level backups (007). It could probably be squashed with 007 if
> > > desired.
> >
> > Adding perldoc to this module should be done separately, let's not mix
> > things. Applying a filter is useful as well to remove for example the
> > contents of pg_xlog, so no objections to it.
>
> Eh, ok. I figured it was so trivial it didn't matter, but will split.

Please don't mess with this one.

> > psql_check sounds wrong to me. I thought first that this triggers a
> > test. Why not psql_simple or psql_basic, or just keep psql.
>
> I guess I'm used to Python's subprocess.check_call so to me it's natural.
>
> I want something that makes it clear that failure is a fatal error
> condition, i.e. "do this in psql and if it produces an error, treat it like
> you would any other error in Perl and die appropriately".

Shrug. psql_check seemed reasonable to me for that.

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