Re: Hacking on PostgreSQL via GIT

Tom Lane wrote:
> Um ... why do either of you feel there's an issue there?
>
> We switched over to $PostgreSQL$ a few years ago specifically to avoid
> creating merge problems for downstream repositories. If there are any
> other keyword expansions left in the source text I'd vote to remove
> them. If you have a problem with $PostgreSQL$, why?

I have to accept the blame for not researching about the repo in the
first place. I didn't know about $PostgreSQL$ - from the looks of it, it
acts _just_ like $Id$. So I guess you use PostgreSQL instead of Id.

As GIT won't touch them, Florian will probably be just fine with his
patches, and I doubt they'll be more than a minor annoyance, if at all.

Keyword expansions are generally bad because SCM tools should track
_content_ - and keyword expansions _modify_ it to add metadata that is
somewhat redundant, obtainable in other ways, and should just not be in
the middle of the _data_. Those modifications lead to patches that have
bogus hunks and sometimes don't apply, MD5/SHA1 checksums that don't
match and a whole lot of uncertainty.

You can't just say "the content is the same" by comparing bytes or SHA1
digests if the committer, the path or the history are different. And it
is a mighty important ability for an SCM.

The argument runs much longer than that - and the flamewars are quite
entertaining. If anyone's keen we're having one right now on
git(at)vger(dot)kernel(dot)org . I am sure Pg hackers will find parallels between
keyword expansion (as a misfeature everyone is used to) and the SQL
travesties that early MySQL is famous for.

I've picked my poison... ran away from MySQL to Pg, and from CVS
/SVN/Arch to GIT. Not looking back :-)

cheers

m
--
-----------------------------------------------------------------------
Martin @ Catalyst .Net .NZ Ltd, PO Box 11-053, Manners St, Wellington
WEB: http://catalyst.net.nz/ PHYS: Level 2, 150-154 Willis St
OFFICE: +64(4)916-7224 UK: 0845 868 5733 ext 7224 MOB: +64(21)364-017
Make things as simple as possible, but no simpler - Einstein
-----------------------------------------------------------------------