Re: Is temporary functions feature official/supported? Found some issues with it.

On Tue, Jan 15, 2019 at 03:36:56PM +0900, Masahiko Sawada wrote:
> Thank you for updating the patch. The patch looks good to me. And the
> new regression test for CREATE EXTENSION seems to work fine but maybe
> it's better to reset client_min_messages at cleanup for safety.

Sure, done. I have been working on this patch more this morning, and
here is a proposal for commit. So please let me know if there are any
objections with that.

9b013dc is the commit which has introduced MyXactFlags, so this means
that we cannot get that back-patched further down than v10. Per the
lack of complains, that's a restriction I can live with.

The first patch is the actual fix to back-patch. The second patch is
an improvement I propose only for HEAD which removes ACCESSEDTEMPREL,
replacing it with ACCESSEDTEMPNAMESPACE. For now I propose to commit
0001, then I'll spawn a new thread to discuss 0002 on -hackers as
XACT_FLAGS_ACCESSEDTEMPREL has the advantage to allow skipping ON
COMMIT DELETE if no temporary tables have been accessed. Perhaps
that's not worth bothering, but that point seems worth poking at, at
least to me.
--
Michael