[PATCH 5/5] Allow to execute non-yielding DDL statements in transactions

[Vladimir Davydov]

On Mon, Jul 08, 2019 at 12:25:06PM +0300, Konstantin Osipov wrote:
> * Vladimir Davydov <vdavydov.dev at gmail.com> [19/07/08 11:45]:
> > > In any case you test a fixed number of combinations, and should
> > > come up with a randomized test. This feature encompasses all DDL
> > > operators, including:
> > > 
> > > - roles
> > > - users
> > > - functions
> > > - SQL triggers
> > > - SQL foreign key constraints
> > > - SQL views
> > > - SQL check constraints
> > > 
> > > Every object should be involved in the test, and the generated
> > > transactions should be random.
> > 
> > Sounds tricky :-/ May be, hand this over to Q&A?
> 
> Nobody is going to write the test better than you. You may even
> hand this over to mermaids or some other mythical creature, if you
> don't want this to happen.

Mermaids are cool ;-)

The problem is such a test is really difficult to come up with, because
we'd have to implement something like a dependency graph describing
which operations can be run after which. IMHO it isn't worth doing just
for the sake of a functional test.

Regarding SQL, the SQL team already has tests dealing with SQL related
objects (like constraints or triggers). Once they introduce support of
transactional DDL, we will get those tests for free.