On Wed, 2008-04-16 at 17:49 -0400, Chris Browne wrote:
> It would presumably not be _completely_ implausible to run a trigger
> when a table was ALTERed; the trouble would be in evaluating the
> semantics what OLD.* and NEW.* ought to contain, _if anything_.
Agreed.
If there was a simple use case and some fairly straightforward
definition of what is needed, that would help move things forward.
For example, would it be useful if the trigger only had access to the
text of the DDL statement? Which DDL statements would it apply to? As
Chris points out, saying "we need everything" just becomes a blocker to
progress.
So if somebody listening would like to research, detail and *justify* a
useful set of additional behaviour then it may be possible to add it.
--
Simon Riggs
2ndQuadrant http://www.2ndQuadrant.com