Re: Doubt about AccessExclusiveLock in ALTER TABLE .. SET ( .. );

On Fri, Jul 31, 2015 at 10:05 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On 31 July 2015 at 02:46, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote:
>
>>
>> > Added. I really don't know if my isolation tests are completely correct,
>> > is
>> > my first time writing this kind of tests.
>>
>> This patch size has increased from 16k to 157k because of the output
>> of the isolation tests you just added.
>
>
> That's too much.

Yes, same opinion as mentioned upthread.

> Why do we need more isolation tests? There isn't anything critical here, its
> just different lock levels for ALTER TABLE. A few normal regression tests
> are fine for this.

Fabrizio went down to 26k with his latest patch by using only a subset
of permutations. To put it shortly, those things are worth testing. We
have the infrastructure to do it, and we lack of coverage in this
area. Hence this patch is a good occasion to do it IMO.
--
Michael