Hi Stephan,
Sounds good. I might be able to help you out with those tests as I'd
love this problem to be fixed. I actually removed all RI and replaced
them with triggers because of the problems I was having. I might be a
bit slow getting back to you each time though as my plate is pretty
full. You can grab me off-list for more info.
Tom.
On Wed, 2002-10-16 at 12:25, Stephan Szabo wrote:
> I'm looking to see if anyone has
> a) a development/testing server they don't mind
> testing very early pre-development stuff on
> b) foreign key systems (using no action) that
> currently have deadlocking or waiting problems
> c) can run realistic requests that have deadlocked
> the current triggers
> d) willingness to help with debugging
>
> I'm in very preliminary stages of trying out the
> dirty read-for update barrier code in a fashion
> that appears to work better in regards to locking
> than what we currently have, but I only have
> simple tests to run from multiple psqls
> interactively right now.
>
> I'm not sure that the no action cases (all I've
> converted at the moment) are working 100% to
> both not lock as badly and still maintain the
> correct semantics and looking for some brave
> souls to help. :)
>
> In any case, I'm still days (at the 1/2 hour
> a day I tend to get to work on it) away from
> having a patch form I'm comfortable with.
>
> If anyone's interested in helping reply here
> or directly. :)
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
--
Thomas O'Dowd - A Noop a day, helps you work rest and play!
tom@nooper.com - http://nooper.co.jp