-----BEGIN PGP SIGNED MESSAGE-----
> I wonder if there's a way to have the server process the matching? That
> way we could have LIKE expressions in the switches, which would be
> simpler in the code and more powerful. I don't know how pg_dump works
> so I can't really answer the question. We desperately need this
> capability however, as patches have been floating since before 8.0 and
> we still don't have it.
It won't fit into the existing code easily, but it could probably be done.
I toyed around with making the regex more robust, but three things
1) The "star at start" and "star at end" catches probably 99% of the cases,
and is way better than what we have now, so better a bird in the hand...
2) It would be a lot more work to send it to the backend or import some
of the regex code.
and most importantly:
3) It would require yet more arguments to pg_dump. The moment we start allowing
regular expression characters that are also valid identifier names (e.g. "."
and "_") we'll need some way to tell pg_dump whether we mean a literal search
or a regular expression one. Which probably means more arguments or at least
modifying the existing one in a possibly nonintuitive, and definitely more
complex, manner. I'm open to suggestions, however, but I don't want to make
things too byzantine for the users.
Greg Sabino Mullane greg(at)endpoint(dot)com greg(at)turnstep(dot)com
PGP Key: 0x14964AC8 200601171718
-----BEGIN PGP SIGNATURE-----
-----END PGP SIGNATURE-----
In response to
pgsql-patches by date
|Next:||From: Tom Lane||Date: 2006-01-17 22:43:24|
|Subject: Re: New pg_dump options: exclude tables/schemas, multiple all, wildcards |
|Previous:||From: Tom Lane||Date: 2006-01-17 22:12:09|
|Subject: Re: Huge number of disk writes after migration to 8.1 |