Performance tests I did with FreeBSD, Linux and PostgreSQL

[Sean Chittenden]

> Please do NOT commit this unless it accompanies a major version
> change!  PGSQL goes to extremes to ensure that you only have to
> dump/reload on major version changes (ie: 7.2.x to 7.3.x), and it'd
> be really bad to surprise people with this.

Absolutely... I guess I posted that bit to the pgsql-hackers list.
I'm pretty sure I'm going to bump things in the -devel port and am
going to talk with Palle regarding this when 7.4 comes out that way it
will, as you state, be along major version bumps.

> I think it would also be a very good idea to see what happens if you
> try and open a database with 8K pages using pgsql compiled for 16K
> pages. If it doesn't detect the error, I don't think it's worth
> doing this; it just seems too risky; though adding a config option
> to the port makefile would be a good alternative in that case.

Talked to Tom, PostgreSQL isn't stupid and does the right thing: it
bombs out before doing any work.

-sc


-- 
Sean Chittenden