On Thu, May 17, 2018 at 10:18:56AM -0400, Tom Lane wrote:
> I think the point you've not addressed is that "syscache callback
> occurred" does not equate to "object was dropped". Can the code
> survive having this occur at any invalidation point?
> (CLOBBER_CACHE_ALWAYS testing would soon expose any fallacy there.)
Thank you for the idea of testing with CLOBBER_CACHE_ALWAYS. I built
postgres with it and run regression tests. I tested both approaches. In
first glance they passed the tests.
There is no concurrent tests for text search feature with two and more
connections. Maybe it would be useful to make such tests. I did it
manually but it is better to have a script.
--
Arthur Zakirov
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company