BUG #17035: assert after commit

PG Bug reporting form <noreply@postgresql.org> writes:

after this commit
https://github.com/postgres/postgres/commit/d18ee6f92d9a22b4fae57f515797b2196bf385c7
I got assert
TRAP: FailedAssertion("portal != NULL", File: "pquery.c", Line: 1758)
in my plugin
https://github.com/RekGRpth/pg_task

I'm inclined to think that means you were doing something you
should not have been doing. But without any details it's
hard to say.

regards, tom lane

Ok, thanks. I only use SPI_execute_plan and before that commit all works fine
with bst regrds, Rek>pth

ср, 26 мая 2021 г. в 18:52, Tom Lane <tgl@sss.pgh.pa.us>:

On 2021-May-26, Tom Lane wrote:

PG Bug reporting form <noreply@postgresql.org> writes:

after this commit
https://github.com/postgres/postgres/commit/d18ee6f92d9a22b4fae57f515797b2196bf385c7
I got assert
TRAP: FailedAssertion("portal != NULL", File: "pquery.c", Line: 1758)
in my plugin
https://github.com/RekGRpth/pg_task

I'm inclined to think that means you were doing something you
should not have been doing. But without any details it's
hard to say.

Hmm, see for example [1]https://github.com/RekGRpth/pg_task/blob/master/work.c#L703 which is doing SPI_prepare_my() [2]https://github.com/RekGRpth/pg_task/blob/master/spi.c and then
SPI_execute_plan_my() ... Does the SPI interface really require that you
create an ActivePortal in the SPI-calling code? This looks a bit
suspicious to me ...

[1]: https://github.com/RekGRpth/pg_task/blob/master/work.c#L703
[2]: https://github.com/RekGRpth/pg_task/blob/master/spi.c

--
�lvaro Herrera Valdivia, Chile

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Hmm, see for example [1] which is doing SPI_prepare_my() [2] and then
SPI_execute_plan_my() ... Does the SPI interface really require that you
create an ActivePortal in the SPI-calling code?

The expectation is that the calling code is already executing inside
some Portal. If it isn't, it's incumbent on the caller to set up
an adequate substitute environment, in particular a transaction
snapshot. The only thing 84f5c2908 changed is that now you get
a guaranteed failure if you neglect to provide that, rather than
failing only in corner cases.

Possibly we should change that Assert to an elog that tries to
make it clear that the blame is on the caller.

regards, tom lane

Ok, I see. But I use SPI_execute_plan in background worker and no
ActivePortal there
with bst regrds, Rek>pth

ср, 26 мая 2021 г. в 22:47, Tom Lane <tgl@sss.pgh.pa.us>:

RekGRpth <rekgrpth@gmail.com> writes:

Ok, I see. But I use SPI_execute_plan in background worker and no
ActivePortal there

Well, you need to provide a snapshot. replication/worker.c
might serve as an example.

regards, tom lane

Ok, thanks. I add PushActiveSnapshot(GetTransactionSnapshot()); after
SPI_start_transaction and PopActiveSnapshot(); before SPI_commit and
it works!
with bst regrds, Rek>pth

чт, 27 мая 2021 г. в 08:22, Tom Lane <tgl@sss.pgh.pa.us>: