sftp bugs

[James Henstridge]

Jan Hudec wrote:

>On Wed, Nov 30, 2005 at 09:33:36 +0800, James Henstridge wrote:
>  
>
>>Matthieu Moy wrote:
>>    
>>
>>>[...]
>>>This is because _parse_url does a
>>>
>>>           # remove leading '/'
>>>           self._path = urllib.unquote(self._path[1:])
>>>
>>>(shouldn't this be a urllib.unquote(self._path.lstrip('/')) to remove
>>>multiple slashes?)
>>> 
>>>
>>>      
>>>
>>I think the URL handling fixes in the patch I posted yesterday should
>>fix this problem.
>>
>>Removing multiple slashes is not necessarily correct though, since the
>>SFTP URI spec is a bit weird: "sftp://localhost/tmp" refers to "~/tmp",
>>while "sftp://localhost/%2Ftmp" refers to "/tmp".
>>
>>It doesn't say how "sftp://localhost//tmp" should be interpreted
>>    
>>
>
>If I read it correctly, it said it's illegal.
>  
>
I'm not sure if that is illegal.  The sftp URI draft spec
(http://www.ietf.org/internet-drafts/draft-ietf-secsh-scp-sftp-ssh-uri-03.txt)
references RFC 3986, which states that the path component can not begin
with two slashes if there is no authority component.  SFTP URIs have a
path component, so that restriction doesn't appear to be valid.

Furthermore, the BNF in the RFC seems to allow the slash to remain
unescaped.

James.