Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Dec 5, 2012 at 7:08 AM, Filip Rembiałkowski
> <filip.rembialkowski@gmail.com> wrote:
>> 3. I made a naive test of simply changing AccessExclusiveLock to
>> ExclusiveLock, and seeing how many regression tests it breaks. It
>> breaks none :-)
> Sure. You could probably downgrade it quite a bit further without
> breaking the regression tests, but that doesn't mean it's safe in all
> cases.
In fact, since the regression tests make no attempt whatsoever to stress
DDL executed concurrently with table accesses, it doesn't prove a darn
thing. The standard regression tests actually try quite hard to avoid
such scenarios, so that the results will be repeatable. You could
perhaps build relevant test cases using the isolationtester
infrastructure, but I don't think anyone has tried particularly.
regards, tom lane