On Tue, Feb 07, 2012 at 07:22:03AM -0600, David Mertens wrote: > The question in my mind is: why is the test for file names on Win32 case > sensitive? Last I checked, Windows doesn't differentiate between "d:" and > "D:"
Win32 also doesn't differentiate between / and \ in pathnames, but if the cache contained \ in pathnames on Win32, it would be "wrong", because the Pod::Html code attempts to store cache names in a normalised fashion with / only. So the test wasn't wrong, per-se, as it was testing that the cache contained the right filenames, and the filenames (expected and actual) would be generated by the same reporting code, so it "should" be getting the case as returned by file system calls. And (I repeat) this test passed on a different Win32 machine, so there's something cooky about how Cwd::cwd() can report "d:" or "D:" on Win32 depending on, well, not sure. It would also be wrong to hard-code "Win32" into the test and do a case insensitive check on it, as VMS is also (generally) case insensitive (and File::Spec knows this). Nicholas Clark