On Mon, Jul 13, 2009 at 7:06 AM, Elliot Shank <[email protected]> wrote:
> David Golden wrote: > >> On Sun, Jul 12, 2009 at 10:49 PM, Elliot Shank <[email protected]> wrote: >> >>> Actually, this begs the question of whether M::B should be producing an >>> error instead of letting T*::Harness deal with it. >>> >> >> Maybe a warning? But the logic is kinda hairy since test_files could be: >> >> * actual file >> * actual directory >> * glob pattern >> >> Maybe if $thing doesn't look like a glob pattern and not -e $thing, then >> warn/die? If recursive isn't set, then even an actual directory that >> exists >> is probably an error. >> > > I was thinking that this would happen after find_test_files() had been > invoked: anything at this point that isn't a plain file would be an error. > On a quick scan of the code, it looks like that will work as M::B resolves globs itself rather than handing them off to Test::Harness. -- David
