On Sat, Jun 23, 2012 at 11:59:00PM -0700, Father Chrysostomos via RT wrote: > So it seems File::stat *is* buggy, and the test is just not good enough > half the time. Or the test is wrong. > > I think the test is wrong, as -l _ will fail after a stat. You have to > do lstat for -l _ to work. Similarly, -l $stat_obj should only return > true if the $stat_obj was returned by File::stat::lstat. For > File::stat::stat, -l should either return false or die. Currently it > returns false. -l _ dies after stat. Yes, I think you're right. I didn't introduce the bug in my recent refactoring of the test, but I didn't spot it either. Thanks for fixing it. Nicholas ClarkThread Previous | Thread Next