> It seems a patch like this was already merged into the upstream of
> this, and got deployed into pgarchives & pglister when I synced those
> up a couple of weeks ago. Sorry about the lateness of that one. Does
> the current version of the code work for what you were looking for?
This is the same as what my patch does, indeed
> It might also be worth considering also checking explicitly for the
> format of PGAUTH_KEY on startup. Something like the attached maybe?
> (against pgweb which is the master of this file) What do you think?
LGTM
--
Célestin Matte