1.0 approaches

[Martin Marques]

Stephen Touset escribió:
> 
> On Feb 8, 2008, at 5:13 PM, Martin Geisler wrote:
> 
>> Stephen Touset <stephen at touset.org> writes:
>>>
>>> This would cause even more confusion, given that no other URI on the
>>> planet uses it. Again, the old way isn't a "convention". It's a
>>> documented standard.
>>
>> And this standard conflicts with the standard use of SSH -- when using a
>> program like Mercurial that runs some data over a SSH connection, then
>> it would be nice if the program followed the standard syntax for SSH
>> paths (like rsync does, for example). I expected Hg to behave like that
>> too in the beginning.
> 
> The fact that scp does not conform to the URI standard is not, in my 
> opinion, an argument to abandon it ourselves. Especially when it's been 
> mentioned before that Mercurial is using a built-in Python URI parsing 
> library, which is (understandingly) incapable of dealing with scp-style 
> paths.

AFAIK the URI definition is for paths that are not relative to the home 
directory of who is doing the shh, as for example, in an http:// 
connection or a ftp:// one.

I always think about what konqueror does when I try to access using 
sftp://. If I add a / at the end it thkes the root of my server, else el 
goes straight to my home directory (rewriting th URI with the home 
directory).

Shouldn't this be the right behaviour?