Bruce Momjian <bruce@momjian.us> wrote:
> Uh, I don't see this patch as applied. Was it not necessary?
No, the bug was reported again and fixed then with: - Force READY portals into FAILED state when a transaction or
subtransactionis aborted Message-Id: <20100218030646.500A37541C5@cvs.postgresql.org> - Fix ExecEvalArrayRef to
passdown the old value of the array element or slice being assigned to Message-Id:
<20100218184147.A9EC97541C5@cvs.postgresql.org>
> ---------------------------------------------------------------------------
>
> Itagaki Takahiro wrote:
> >
> > Robert Haas <robertmhaas@gmail.com> wrote:
> >
> > > It looks like this is enough to reproduce the cache lookup failure:
> >
> > The "cache loopup failure" part could be fixed by the attached patch.
> > It forbids explaining if the transaction is in error state.
> >
> > I cannot reproduce "unexpected refassgnexpr" and "unexpected FieldStore"
> > errors yet. We might need another fix for them.
Regards,
---
Takahiro Itagaki
NTT Open Source Software Center