[PATCH] test_timestamp.py changes for win32
bialix at ukr.net
Thu Apr 19 18:12:51 BST 2007
-----BEGIN PGP SIGNED MESSAGE-----
John Arbash Meinel пишет:
> I'm pretty sure Martin's patch supersedes this one.
Yes, I mark it accordingly.
>But I wanted to
> point out one thing.
> I think the proper way of doing features is to turn them into
> singletons. By creating:
> class _GmtimeNegative...
> GmtimeNegativeSecondsFeature = _GmtimeNegative...()
> The importance of this is that a given Feature only checks if it is
> supported 1 time. Rather than checking every time it is needed.
OK. I understand. Will do.
> This is especially important for things that start to get expensive
> (like spawning strace to see if it is available). By using a singleton
> we only have to check 1 time regardless if there are 50 tests that need
> that feature.
> I'd really like to see the test suite updated so that the Transport
> tests use features rather than not returning test cases. So you could
> see that 500 tests are skipped because of not having paramiko, and 200
> are skipped because of not having 'medusa'.
And another couple of hundreds because of symlink lacks on win32? ;-)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the bazaar