Hiroshi Inoue <[EMAIL PROTECTED]> writes:
> There was a bug report in Japan that plpgsql
> crashes if select returns no row. This seems
> a bug introduced by your latest change on
> pl_exec.c.
Indeed. Too bad that report didn't arrive on Friday.
I am mightily embarrassed :-(
Tom Lane wrote:
>
> Robert Hentosh <[EMAIL PROTECTED]> writes:
> > I then proceed to load the OpenACS datamodel and had the backend crash.
> > This datamodel loads fine on 7.1.
>
> Ugh.
>
There was a bug report in Japan that plpgsql
crashes if select returns no row. This seems
a bug introduced
Robert Hentosh <[EMAIL PROTECTED]> writes:
> I then proceed to load the OpenACS datamodel and had the backend crash.
> This datamodel loads fine on 7.1.
Sigh. Looks like I managed to break plpgsql's handling of SELECT with
no data found. Mea maxima culpa (though the regression tests perhaps
des
Robert Hentosh <[EMAIL PROTECTED]> writes:
> I then proceed to load the OpenACS datamodel and had the backend crash.
> This datamodel loads fine on 7.1.
Ugh.
> I can send the datamodel and core file if needed.
Datamodel please, corefile no.
regards, tom lane
--
I just checked out CVS this morning the REL7_1_STABLE branch. I
configured it with:
./configure --enable-debug
and ran the regression test fine on OpenBSD 2.8 (AMD processor) (The
same problem has been reproduced by someone else on RH6.2)
I then proceed to load the OpenACS datamodel and
