[MERGE] test framework distinguishes skips

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

John Arbash Meinel wrote:
> Aaron Bentley wrote:

> Yeah, but since the pqm runs with LANG=C (make test), all unicode tests
> are skipped. So I don't think we want to look through 200 skipped tests
> to see that one was skipped for the wrong reason.

As I noted, I've already reverted to the exception handling style we had
before.  Is there something more that you think we should be doing?

And when we're listing skipped tests, I hope to group them by reason, e.g.

210 tests skipped because of platform deficiencies
200 tests skipped due to "Non-unicode filesystem"
10 tests skipped due to "No Symlink support"
Tests skipped due to "Non-unicode filesystem":
  test_urlutils.TestUrlToPath.test_posix_local_path_to_url
  test_urlutils.TestUrlToPath.test_win32_local_path_to_url

etc.

>>The symlink and file mode checks are done with TestCase.require_symlinks
>>and TestCase.require_file_modes

> It seems reasonable to have those on TestCaseInTempDir, rather than
> TestCase. Simply because you don't really require either one if you
> aren't writing to the filesystem.

I don't know if that's really true.  You might be using the
(memory-backed) NTFSTransport, for example.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFErn7D0F+nu1YWqI0RAuBdAJ4oYBJrpdKKiuG5/2i8Rec5lRm6+ACgh6BI
OLgetvjCa93oEGEGD6vhEBU=
=E0SF
-----END PGP SIGNATURE-----