Message88050
| Author |
loewis |
| Recipients |
giampaolo.rodola, jafo, jaraco, lemburg, loewis, nnorwitz, swarren |
| Date |
2009年05月18日.20:29:07 |
| SpamBayes Score |
8.199372e-10 |
| Marked as misclassified |
No |
| Message-id |
<4A11C511.8040304@v.loewis.de> |
| In-reply-to |
<1242651062.23.0.881837774389.issue1578269@psf.upfronthosting.co.za> |
| Content |
> I would appreciate if someone could review the patch and comment on the
> technique.
There are a few minor issues; overall, it looks correct:
- the test for "this is windows" should just use MS_WINDOWS.
- don't declare variables in the middle of a block; we use C89.
- the argument parsing looks incorrect; take a look at rename()
for guidance. Supporting bytes is optional; IMO, requiring
Unicode strings for the API would be fine.
I don't quite understand why you can't implement lstat, or why
you would want to call into ntpath. Take a look at how stat()
is implemented in the nt module. |
|