Re: Dynamic shared memory areas

On Thu, Dec 1, 2016 at 6:33 AM, Thomas Munro
<thomas(dot)munro(at)enterprisedb(dot)com> wrote:
> Please find attached dsa-v8.patch, and also a small test module for
> running random allocate/free exercises and dumping the internal
> allocator state.

OK, I've committed the main patch. As far as test-dsa.patch, can we
tie that into make check-world so that committing it delivers some
buildfarm coverage for this code? Of course, the test settings would
have to be fairly conservative given that some buildfarm machines have
very limited resources, but it still seems worth doing. test_shm_mq
might provide some useful precedent.

Note that you don't need the prototype if you've already used
PG_FUNCTION_INFO_V1.

I'm not sure that using the same random seed every time is a good
idea. Maybe you should provide a way to set the seed as part of
starting the test, or to not do that (pass NULL?) and then elog(LOG,
...) the seed that's chosen. Then if the BF crashes, we can see what
seed was in use for that particular test.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company