On Mon, 02 Sep 2002 01:20:51 -0400, Tom Lane wrote:
>
> My two cents: once we ship beta1 we should try really really hard to
> avoid forcing an initdb cycle before final release. We can make all the
> portability fixes and code fixes we like, but we have to avoid
> disk-file-contents changes and system catalog changes. If we force an
> initdb then we're penalizing beta testers who might have been foolish
> enough to load large databases into the beta version --- and yeah, there
> were no guarantees, but will they do it again next beta cycle?
Well, speaking as a user and tester, I don't mind an initdb between beta
releases -- and I'm planning on loading ~8GB of data. I've been playing
with the CVS tip all week using a ~2GB dataset, and having to initdb
every other day or so. No biggie, and I'm getting to test pg_dump, as a
bonus! ;-)
I suppose some delicate souls would be put off by an initdb, or get sore
at experiencing the reason Beta != Production, but I think the serious
users will prefer that you developers have a free hand to do whatever is
necessary to make 7.3 as good as it can be.
My US$0.02,
Gordon.
--
"Far and away the best prize that life has to offer is the chance to work hard at work worth doing." -- Theodore
Roosevelt