Re: [WIP] UNNEST(REFCURSOR): allowing SELECT to consume data from a REFCURSOR

> On 11 Jan 2020, at 12:04, Dent John <denty(at)QQdd(dot)eu> wrote:
>
>> On 10 Jan 2020, at 15:45, Daniel Verite <daniel(at)manitou-mail(dot)org> wrote:
>>
>> postgres=# select unnest('c'::refcursor);
>> server closed the connection unexpectedly
>> This probably means the server terminated abnormally
>> before or while processing the request.
>> The connection to the server was lost. Attempting reset: Failed.
>
> Okay. That’s pretty bad, isn’t it.

I’ve addressed the issue, which was due to me allocating the TupleDesc in the multi_call_memory_ctx, which seemed quite reasonable, but it actually needs to be in ecxt_per_query_memory. It seems tSRF-mode queries are much more sensitive to the misstep.

A v4 patch is attached, which also renames UNNEST(REFCURSOR) to ROWS_IN(REFCURSOR), adds a test case for use in tSRF mode, and makes some minor fixes to the support function.

I have not yet made steps towards documentation, nor yet rebased, so the Makefile chunk will probably still fail.

denty.