On Thu, Sep 05, 2019 at 09:59:51AM -0400, Alvaro Herrera wrote:
> Patch looks good to me, please push.
Thanks, applied without the calls to set_pglocale_pgservice().
> Generally speaking I find the 'progname' handling a bit silly (since we
> have it both as a variable in each program and also in logging.c
> separately), but that's not the fault of this patch, and this patch
> doesn't make it worse. That said, I think some other messages in
> vacuumlo can be made pg_log_somelevel(), based on occurrences of
> 'progname'.
That would mean moving some of the messages from stdout to stderr. We
could do that.
--
Michael