On Sun, Sep 10, 2017 at 6:25 PM, Stephen Frost <sfr...@snowman.net> wrote: > * Michael Paquier (michael.paqu...@gmail.com) wrote: >> As there begins to be many switches of this kind and much code >> duplication, I think that some refactoring into a more generic switch >> infrastructure would be nicer. > > I have been thinking about this also and agree that it would be nicer, > but then these options would just become "shorthand" for the generic > switch.
I don't really like the "generic switch infrastructure" concept. I think it will make specifying behaviors harder without any corresponding benefit. There are quite a few --no-xxx switches now but the total number of them that we could conceivably end up with doesn't seem to be a lot bigger than what we have already. Now, if we want switches to exclude a certain kind of object (e.g. table, function, text search configuration) from the backup altogether, that should be done in some generic way, like --exclude-object-type=table. But that's not what this is about. This is about excluding a certain kind of property (comments, in this case) from being backed up. And an individual kind of object doesn't have many more properties than what we already handle. So I think this is just an excuse for turning --no-security-labels into --no-object-property=security-label. To me, that's just plain worse. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers