Thomas G. Lockhart wrote:
>
> > Has anyone tried building the backend with a debugging malloc library?
>
> This has been done in the past by a couple of people who had access to
> Purify. I recall that it helped find several problems, but also
> generated quite a few spurious messages because of the way we
> allocate/deallocate blocks of memory during queries.
>
> Our current best debugging technique is for a bunch of us to get all in
> a froth about a problem until Vadim gets fed up with our flailing around
> and fixes it himself ;-)
Thanks, Tom -:))
But it seems that Purify (etc) would be very usefull
in pg_user bug. (Imho, Purify could also help with
indexing bug!)
I finally compiled 6.4-current on SPARC/Solaris 2.5.1
(gcc 2.7.2.1) but unfortunately I can't reproduce
pg_user bug here -:(
So, I would like to switch to primary/foreign/unique key
infos...
Vadim