Robert - can you please comment on what you are willing to commit in order to close out your open item here. My take is that the design for this, the tabular form a couple of emails ago (copied here), is ready-to-commit, just needing the actual (trivial) code changes to be made to accomplish it.
That said, from a readability standpoint, it was easier for me to follow
the tabular form vs. the sentence form.
May be possible to reach a agreement on the sentence form. Similar descriptions used
for referential constraints in the \d command:
I think we should consider the tabular form with translatable headers to be the acceptable choice here. I don't see enough value in the sentence form to make it worth trying to overcome the i18n objection there.
As for tabular form it looks more natural to have a separate psql command
for pg_auth_members system catalog. Something based on this query:But is it worth inventing a new psql command for this?
IMO, no. I'd much rather read "admin, inherit, set" than deal with true/false in columns. I think the newlines are better compared to repetition of the rolname as well.
I'm also strongly in favor of explicitly writing out the word "empty" instead of leaving the column blank for the case that no options are marked true. But it isn't a show-stopper for me.