On 2/1/16 4:19 PM, Tom Lane wrote:
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
>> .... Anyway I think the tests here are
>> massive and the code is not; perhaps people get the mistaken impression
>> that this is a huge amount of code which scares them. Perhaps you could
>> split it up in (1) code and (2) tests, which wouldn't achieve any
>> technical benefit but would offer some psychological comfort to
>> potential reviewers. You know it's all psychology in these parts.
>
> Perhaps the tests could be made less bulky. We do not need massive
> permanent regression tests for a single feature, IMO.
I'd certainly like to but pgaudit uses a lot of different techniques to
log various commands and there are a number of GUCs. Each test provides
coverage for a different code path.
I'm sure they could be reorganized and tightened up a but I don't think
by a whole lot.
--
-David
david@pgmasters.net