ubuntuone-users team mailing list archive
-
ubuntuone-users team
-
Mailing list archive
-
Message #00772
Re: [REST API] Chars to exclude from percent encoding
Hi, I asked around a little :-)
URL encoding for oauth is tricky, because it's not exactly the usual URL
encoding. It's defined in
http://tools.ietf.org/html/rfc5849#section-3.6but we are using a
library in the server to handle it, so I am not
absolutely certain we are doing it exactly as the rfc requires.
It may be that your client library is expecting the unencoded URL and does
its own encoding?
On Thu, Jan 9, 2014 at 3:27 PM, José Expósito <jose.exposito89@xxxxxxxxx>wrote:
> Hi list,
>
> From a Qt client, I'm encoding the URL to be able to make some file
> operations:
>
> https://github.com/JoseExposito/ubuntuone-qt-files/blob/master/src/com/egg-software/ubuntuone-qt-files/network/AbstractMessage.cpp#L67
>
> For example, to download a file named "(HI).txt" I send a request like
> this:
>
> https://files.one.ubuntu.com/content/~/Ubuntu%20One/%28HI%29.txt
>
> But I have noticed that the Ubuntu One server returns a "Host requires
> authentication" error. To avoid this error I have excluded the chars "("
> and ")" from the encoding, and all work like a charm with a URL like this:
>
> https://files.one.ubuntu.com/content/~/Ubuntu%20One/(HI).txt
>
> My question is.. Where can I get the full list of chars to exclude from
> the encoding?
>
> _______________________________________________
> Mailing list: https://launchpad.net/~ubuntuone-users
> Post to : ubuntuone-users@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~ubuntuone-users
> More help : https://help.launchpad.net/ListHelp
>
>
Follow ups
References