postgres changes

[Matthew Vanecek]

On Mon, 2003-08-11 at 21:18, Linas Vepstas wrote:
> OK, I reviewed the rest of the postgres changes, they looked innocuous.
> I fixed one minor thing in table-version.sql, the version number should be
> 1.5.1 not 1.5.0 (the 1.5.0 is written before the changes are started, 
> then 1.5.1 to signify they're complete.  Just in case things crashed 
> in the middle, ... )
> 
> 
> I actually looked at newtables.h  Wow! that's a lot of work!
> 

Thanks.  Took a bit of testing to get it just right, but I wanted it to
be as close to portable and complete as possible (too bad ALTER...RENAME
isn't an SQL standard--won't be able to do that with libdbi. :( ).

> 

Gar, the /. effect on CVS!!!!  The Postgresbackend change is needed to
support the internalization of transactions in
upgrade.c/add_timezone..().  Accordingly, I transactionalized the other
functions in upgrade.c, too, so the pgendUpgradeDB really needs to go
back where I had it, and the transactionalizations re-applied to the
upgrade functions.

I did do the Entry->Split change in upgrade.c and didn't even think
about it until after I sent the patch. Gar!!  Would've sent something in
earlier, but dinner called...other than that, the other patches should
stay the same.

I didn't realize what the revision in the version string was used for. 
I'll remember that for the future.

> --linas
-- 
Matthew Vanecek
perl -e 'print $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'
********************************************************************************
For 93 million miles, there is nothing between the sun and my shadow except me.
I'm always getting in the way of something...