On Aug 1, 2007, at 10:56 AM, Richard Huxton wrote:
You could write a small cron-script that dumped the schema once
every 5 minutes so it could be picked up by svn.
I think most people have a separate collection of schema-creation/
update scripts that they keep under version control. All changes
are then through running these.
You would have to do it via polling, since schema changes cause no
events to be generated (ie, you can't attach a trigger to a schema
change.)
But the *right* way is to make schema change scripts as "delta"
files, add them to your repo, test them on your staging environment,
then apply them to your production environment. That way you can
reconstruct your DB at any time and *know* it will work.
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
