Greg Hudson <ghudson@MIT.EDU> writes:
> > I'm sure we agree that it would be slightly nicer to be using a PRNG
> > under our own control, but unless you've got one lying around
> > somewhere, I don't think it's worth the hassle until we have a
> > problem.
>
> rand() is just "r = r * 1103515245 + 12345". Not exactly rocket
> science. It's not a very good random number generator, but we don't
> need an especially good one for this application.
>
> I will go ahead and make random-test.c stop using rand(), just to set
> people's minds at ease. I'll also see about integrating it properly
> with the test suite and having it run multiple times.
Sounds great.
The current testing framework is designed for run-it-once regression
testing. I would guess that random testing would benefit more from a
framework that ran the tests constantly on some machine somewhere, and
checked out new sources once a day or once an hour or something.
There's an idle machine in the Chicago office (they got it for me to
use, but I'm usually in Bloomington); perhaps we could hack something
up there.
Received on Sat Oct 21 14:36:09 2006