"Erik Rijkers" <er(at)xs4all(dot)nl> writes:
> I ran into a bug in pg_restore, where it silently omits
> to restore a constraint (here a primary key):
pg_restore's notion of what -t means is very much more restricted than
pg_dump's is. We might change this someday, but don't hold your breath
(ie, it's a missing feature not a bug).
In the example at hand I'd suggest just using the -t switch on the dump
side and not passing anything to pg_restore.
regards, tom lane