Re: PL/pgSQL

On Wednesday 31 March 2004 18:58, Bob(dot)Henkel(at)hartfordlife(dot)com wrote:
> Below is the info from the on-line docs for 7.4.2. Can someone tell me how
> to hook into the error mechanism to notice that this happens. Like the doc
> says you can.

If what you want to do is detect errors and work around them, this will
require nested transactions.

> If PL/pgSQL had that I would be 100% happy with PostgreSQL instead of 99%
> happy(not that anyone cares about my happiness).

The core care about all our happiness :-)
But there aren't that many of them :-(
But I do know there's interest in getting nested transactions working :-)
But it's not ready yet :-(
But you can help :-)

> That's why I'm willing to
> do something about this if I can only get some help and guidance on doing
> this. Any coders out there want to help me out on figuring this out or
> adding this to the PostgreSQL source for all to use and enjoy?

Firstly - good for you. People like you are the reason PG is such a pleasure
to work with.

Secondly, check the pgsql-hackers list (start with the archives). I don't know
what the state of nested-transactions is, or whether you have enough C
knowledge to help out at this stage.

If you can't help with the code, but you can sync CVS and compile from source,
then you can help testing.

If you don't feel comfortable with that, then there will need to be
documentation written (later) and examples, and articles that can go on
postgresql.org or news sites.

> To me this
> is always that one thing that PostgerSQL doesn't have that I think is not a
> plus but a needed piece of functionality.

I'm sure you'll be able to move this feature along at some point - if like me,
you lack the C skills to code, you may have to be patient though.

--
Richard Huxton
Archonet Ltd