Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Well, I thought I did it, and it did work on my limited number of test
> > cases. Seems you got it fully working.
>
> Actually, it failed for me (and evidently for Barry) on exactly the test
> case you posted along with the patch. You said
>
> > test=> set autocommit = off;
> > SET
> > test=> commit;
> > WARNING: COMMIT: no transaction in progress
> > COMMIT
>
> but in fact I saw the COMMIT succeeding without complaint. I was
> meaning to ask you just what code you'd tested, because this morning's
> CVS tip did *not* behave as above.
I am stumped myself as well. I still have the CVS of my old code, and
it fails just as you saw, but I know I tested it and copied that into
the email via cut/paste so my only guess is that I tweaked something
after I ran the test and if broke something else. If you got it all
working now, I won't research further.
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square,
Pennsylvania19073
