On Tue, May 08, 2007 at 10:51:52PM +0100, David Cantrell wrote:
> Tim Bunce wrote:
>
> >So, David, which version of Math::BigInt are you using?
> >
> >According to the change notes for Math::BigInt 1.86 there should
> >be test failures for 1.85 on 32bit systems. If you're using 1.85 did you
> >see those and install it anyway?
>
> With no Math::BigInt installed at all, DBI 1.55 passes its tests.
>
> With Math::BigInt 1.83, DBI passes.
> With Math::BigInt 1.86, DBI passes.
>
> With Math::BigInt 1.84 and 1.85, Math::BigInt's tests seem to get stuck,
> so I interrupted them with C-c, which makes CPAN.pm think the test
> result is 'unknown' instead of 'fail', so it goes on to install the
> broken module then DBI fails.
Great. Just what I needed. Does the patch below fix it for those cases?
Tim.
Index: t/01basics.t
===================================================================
--- t/01basics.t (revision 9503)
+++ t/01basics.t (working copy)
@@ -231,6 +231,8 @@
SKIP: {
skip("Math::BigInt < 1.56",2)
if $DBI::PurePerl && !eval { require Math::BigInt; require_version
Math::BigInt 1.56 };
+ skip("Math::BigInt $Math::BigInt::VERSION broken",2)
+ if $DBI::PurePerl && $Math::BigInt::VERSION =~ /^1\.5[45]/;
cmp_ok(DBI::hash("foo1",1), '==', -1263462440, '... should be -1263462440');
cmp_ok(DBI::hash("foo2",1), '==', -1263462437, '... should be -1263462437');
}
