On Mon, Sep 03, 2001 at 11:31:35AM +0200, H. Merijn Brand wrote:
> On Mon 03 Sep 2001 10:42, Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > What do we do about t/op/rand.t?
> > It's expected to fail .1% of the time. We're running enough tests that this
> > should show up regularly.
Why not just increase the # of reps by a factor of ten? That should
lover the chanced of the test failing. It increases the time of the
test by less than a half second here (PowerPC G3/266).
I'll do that. And strip that useless srand.
> And the same for time.t, the test that sparsly fails in Simon's
> reports, though I think someone is looking at loosening the
> restraints for that test to pass as by Arthur's request.
I'm fixing up that test. What bit tends to fail?
--
Michael G. Schwern <[EMAIL PROTECTED]> http://www.pobox.com/~schwern/
Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One
some sniff tubes of glue,
but I prefer to rub it in my poo,
and read Whinny and Tigger, too
like ODB at the Brooklyn Zoo.
-- Ubergirl's beau
