Closed Doomas closed 9 years ago
I think you're right that this should change, but I think latin-1
is the encoding that we should be passing, maybe.
There is a bug, but I think we want native_(value, 'latin-1')
instead. This is what WebOb does internally, because that's what WSGI servers expect.
With text_
we get unicode
type on Py2. With native_
we get str
on both, which fits PEP3333's language about unicode and what WebOb seems to have discovered is best for compatibility with servers.
Yes, you're right, native_
with latin-1
is the correct choice.
If you could make that change, squash the commits, and force push that up to this branch I can merge this with that change. That would be excellent. Thank you!
GitHub doesn't notify me when you push commits on a PR, only when you comment. But I noticed you made the change. Thanks.
If you wouldn't mind squashing the commits, that would be beautiful and then I'll hit merge.
Sorry, my bad. Here you go.
Thanks so much!
I ran into some Python 3 compatibility issues. This fix should be also compatible with Python 2.7+
Waitress 0.8.10 excepts strings instead of bytes.