On Fri, Jun 24, 2005 at 09:11:53AM +0100, Steve Hay wrote: > In the absence of a bleadperl smoke (until I or someone else finds time > to fix Test-Smoke for the new harness output), here's a new test failure > using a default threaded config on Win32/VC++: > > op/gv.t > not ok 67 - DESTROY was called > # Failed at ./test.pl line 249 > # got '' > # expected /(?-xism:^Farewell M=SCALAR)/ > > I don't think this was happening yesterday, so its probably something in > changes 24962-24968.
I somehow doubt it, as the test was added last night. http://public.activestate.com/cgi-bin/perlbrowse?patch=24966 Is it something to do with test.pl not capturing the outuput? If you run the test program on its own, what do you get? package M; sub DESTROY {warn "Farewell $_[0]"} package main; bless \$A::B, 'M'; *A:: = \*B::; __END__ Nicholas Clark
