usage of unicode_literals breaks "import *" about pyjwt

[mark-adams]

Which specific version of Python does that happen in?

[mark-adams]

Nevermind, I got it in 2.7.

[wbolster]

The source of the problem (for Python 2) is that the unicode_literals future import makes it impossible to use "native strings" in the module, and since "native strings" are used all over the place in Python, this causes all kinds of weird errors.

[mark-adams]

I'm also going to go ahead and mention the article you linked to in #67, since it seems to provide a good discussion on the issue: PythonCharmers/python-future#22

[jpadilla]

Anything happen when we remove unicode_literals?

[wbolster]

Don't think so, since the API doesn't have any str/unicode default args or return values, afaics. Perhaps the dict lookups for the algos (that was reworked anyway) needs some special care so that the caller can always specify a native string for those.

[jpadilla]

@wbolster nothing seemed to happen in #75.

[jpadilla]

Closed via 6a0f184

[wbolster]

for reference, this was fixed via #76

[merwok]

For the record, one can use the b'string of bytes' syntax in 2.6+ to have “native strings” in a module with unicode literals enabled.

[wbolster]

No, because it would result in bytes on Python 3, and those are not native strings.

"Éric Araujo" [email protected] schreef op 23 maart 2015 20:28:29 CET:

For the record, one can use the b'string of bytes' syntax in 2.6+
to have “native strings” in a module with unicode literals enabled.

---

Reply to this email directly or view it on GitHub:
#73 (comment)

Sent from my Android device with K-9 Mail. Please excuse my brevity.

[mark-adams]

Isn't this resolved already? 😄