Tels wrote:
> > I had the same problem on Debian an x86 with 20002, 20006 and 20008
> > when running 'make test'. I ran the test with 'harness' hundred times
> > but it always succeeded. Strange. I increased my stacksize to 64M, no
> > change.
Note that "make test" runs t/TEST, while "make test_harness" runs
t/harness. Before applying Tels' upgrade, I tested it with t/harness,
not t/TEST --
> I think it is because of this output:
>
> ok 2119
> start 8 divider 2 up 2
> at 1 (0 0)
> at 1 (1 8
> at 4 (1 0)
> at 4 (4 0
> at 2 (4 1)
> at 2 (2 2
> 8 org 8
> ok 2120
>
> That might confuse "make test", since there are only 2648 tests.
I applied this :
Change 20010 by [EMAIL PROTECTED] on 2003/07/05 12:06:41
Comment out some debugging output in Math::BigInt, that
causes confusion in t/TEST when running Math/BigInt/t/bare_mbi.t.
Affected files ...
... //depot/perl/lib/Math/BigInt.pm#52 edit
Differences ...
==== //depot/perl/lib/Math/BigInt.pm#52 (text) ====
@@ -2188,22 +2188,22 @@
#my $lastlast = $x+$y;
my $divider = $self->new(2);
my $up = $y-1;
- print "start $org divider $divider up $up\n";
+ #print "start $org divider $divider up $up\n";
while ($last != $x && $lastlast != $x)
{
- print "at $x ($last $lastlast)\n";
+ #print "at $x ($last $lastlast)\n";
$lastlast = $last; $last = $x->copy();
- print "at $x ($last ",($org / ($x ** $up)),"\n";
+ #print "at $x ($last ",($org / ($x ** $up)),"\n";
$x->badd($org / ($x ** 2));
$x->bdiv($divider);
}
- print $x ** $y," org ",$org,"\n";
+ #print $x ** $y," org ",$org,"\n";
# correct overshot
while ($x ** $y < $org)
{
- print "correcting $x to ";
+ #print "correcting $x to ";
$x->binc();
- print "$x ( $x ** $y == ",$x ** $y,")\n";
+ #print "$x ( $x ** $y == ",$x ** $y,")\n";
}
$x->round(@r);
}
