Re: enable/disable trigger (Re: Fwd: [HACKERS] Open items)

Thanks for comments, Neil.

Some are fixed.

Neil Conway wrote:
> Also, you should probably skip the simple_heap_update() if the user
> tries to disable an already-disabled trigger, to avoid pointless MVCC
> bloat (and same for enabling an already-enabled trigger, of course).

Do we need some log/warning message for the user in these cases?

> If someone pg_dump's a table with a disabled trigger, should the dump
> enable or disable the trigger? I'd be inclined to say pg_dump should
> preserve the state of the database it is dumping, and so the trigger
> should be disabled. In that case I believe pg_dump needs to be updated.

I think so too.

> The patch needs to update the documentation and add regression tests.
> psql tab completion might also be worth adding.

I'll dive into pg_dump and psql this weekend...

--
NAGAYASU Satoshi <nagayasus(at)nttdata(dot)co(dot)jp>