On Fri, Aug 29, 2014 at 4:34 AM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Robert Haas <robertmh...@gmail.com> writes: >> On Thu, Aug 28, 2014 at 12:22 PM, Fujii Masao <masao.fu...@gmail.com> wrote: >>> Another approach is to change pg_dump so that it encloses the relopt >>> values with single quotes. This is the same approach as what >>> pg_dumpall does for database or role-specific settings. > >> To me, this doesn't seem nearly important enough to justify breaking >> pg_dump compatibility. AAUI, this is just a cosmetic improvement, so >> we shouldn't break functional things for that. > > Indeed. I am not convinced that pg_dump is the only client-side code > that would get broken, either. > >>> Further other approach is to change the reloptions code so that it >>> always stores the plain value without the units (i.e., 1000 is stored >>> if 1s is specified in autovacuum_vacuum_cost_delay)in pg_class. > >> This seems like the way to go.
Agreed. > > Yeah, it's the best idea I can think of either. It's a tad annoying but > I think we don't want to take the compatibility risks of storing > unit-ified values in reloptions. > > In the meantime, the buildfarm is still all red. Can we please revert > this patch until a fixed version is ready? OK, reverted. Regards, -- Fujii Masao -- Sent via pgsql-committers mailing list (pgsql-committers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-committers
