Take your time, no problem. The run-time is pretty acceptable now (less than
1 minute), even if, tests for NH2583, represent ~25% of the time needed to
run all others ~4000 tests.

The discussion around NH2583 is interesting and it goes deep in the problem.
I have appreciate all the effort you and Patrick have put in it however we
have to think in the future of those tests. I can't explain which is my
feeling when something break in NHibernate.Test.Legacy and some other areas;
I would try to avoid the same error especially for a complex area as our
LINQ provider, nothing more than that, really.


On Tue, Apr 26, 2011 at 9:30 AM, Harald Mueller <[email protected]>wrote:

> Fabio -
>
> I don't need 4K entries - but what I tried to say is that "I" (we?) need
> full input coverage for critical functions *or* a reliable method that
> reduces the input (not some guesswork like "pair-wise coverage" or even
> manually selecting "just a few records" or the like).
> I want to pay *all* attention to the runtime of tests! - it just did not
> occur to me - or Patrick, as far as I know.
> We'll try to find a solution - maybe there's a simple one! Anyone has
> suggestions?!? It might, however, take "a little" - a few days ok?
>
> Regards
> Harald
>
> -------- Original-Nachricht --------
> > Datum: Tue, 26 Apr 2011 08:31:46 -0300
> > Von: Fabio Maulo <[email protected]>
> > An: "[email protected]" <
> [email protected]>
> > Betreff: Re: [nhibernate-development] Re: My feeling of our Tests
>
> > Harald,
> > I know that this is part of the cost/gain but please say me that you
> > really need 4K rows to test somethig in a method.
> > btw, I have reduced the time-to-run of those tests and I realy hope
> > I'll never see somethig broken there; what I'm asking is to try to pay
> > a bit of attention to check if we really need all those rows in the DB
> > and some comment here and there (in the test) to understand the
> > scenario tested by a method.
> >
> > Herald, the time to run all NH's tests, for me, is important (if you
> > have a look to the SVN log I'm sure you will understand).
> >
> > --
> > Fabio Maulo
> >
> > P.S. I have changed the logic of the SQL generated for
> > binary-equality, when you have time please have a look to the result,
> > thanks.
> >
> >
>
> --
> NEU: FreePhone - kostenlos mobil telefonieren und surfen!
> Jetzt informieren: http://www.gmx.net/de/go/freephone
>



-- 
Fabio Maulo

Reply via email to