Strips new lines in fetch_token response body

[tommarmstrong]

This change strips new line characters from the response body when requesting access tokens from an oauth server. This solves an error I was encountering when trying to authenticate against the WordPress API which returns a new line after the headers and before the URI-encoded response.

This is the format specified in Section 2.3 of RFC 5849 however validation of the response was failing as \r and \n are not safe characters as defined by oauthlib. This caused an exception to be thrown when calling oauthlib's urldecode function on line 30 of oauth1_session.py. By striping those characters before calling urldecode, the presence of a new line in the response body does not cause an exception to be thrown.

[coveralls]

Coverage remained the same at 86.758% when pulling 9159b57 on tommarmstrong:master into 244e835 on requests:master.

[Lukasa]

One HTTP line delimiter (CRLF) should have been fine, because that's used as the delimiter for the end of the header block. However, your bug seems to be caused by Wordpress adding an extra line delimiter, which is factually part of the request body.

While they really shouldn't do that, I see no reason not to strip whitespace from the token body here.

I freaking hate OAuth. 😭

[tommarmstrong]

Hahaha, relatable. Thanks for merging :)

[tommarmstrong]

Hi @Lukasa, when do you think this will make it into a release? Thanks

[Lukasa]

Hrm, probably sometime soon. I'll add a TODO to release early next week.

[tommarmstrong]

Thanks :)

…

On Thu, 9 Feb 2017 at 20:29 Cory Benfield ***@***.***> wrote: Hrm, probably sometime soon. I'll add a TODO to release early next week. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#255 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACmQtLLtQv8beIM1xaN4VN38-XUHlPSnks5ratxtgaJpZM4KWeMc> .