segfault if SETOF RECORD-returning function used without AS at callsite about pljava HOT 5 CLOSED

What do you mean by "everybody gets kicked out of PostgreSQL"? I would assume that only the process corresponding to the connection that experience this problem would segfault and that other connections are unaffected.

from pljava.

What gets logged for everybody else is:

The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
terminating connection because of crash of another server process

After everybody's kicked out, the postmaster goes through a recovery scan of the database files, and then it is back open for business and users can reconnect.

from pljava.

If you're able to build and test this fix, I'd appreciate if you could confirm that it works OK.

from pljava.

Hi,

It might be a little while - eventually I'll move to the github
sources and play with the maven build in a test database instance,
but that's not what's on the table this week. I did look at the change
and it's the same change I would have made, so if you were able to
reproduce the segfault before the change and not after, I'm willing
to be convinced it's fixed. :)

Thanks for taking care of that. Once I learn to navigate the
paperwork on this end, I might reach a point where I can just submit
a pull request with less than a week of bureaucratic exercise.

Regards,
-Chap

On 02/22/2013 08:34 PM, Thomas Hallgren wrote:

If you're able to build and test this fix, I'd appreciate if you could confirm that it works OK.

---

Reply to this email directly or view it on GitHub:
#7 (comment)

from pljava.

I've now had a chance to build and test, and now I do get the new message instead of the segfault. Thanks!

I edited my original posting to include one detail I had at first left out that I didn't realize was necessary to replicate the problem ... sorry about that.

-Chap

from pljava.