Paths and URLs in Branch.base ?

[John A Meinel]

Martin Pool wrote:
> On 25 Nov 2005, Robert Collins <robertc at robertcollins.net> wrote:
>> On Thu, 2005-11-24 at 10:39 -0500, Aaron Bentley wrote:
> 
>>> Alexander Belchenko wrote:
>>>>     * Patch #25. [Win32-specific] Default value for nick use '/' on
>>>> Windows. Changed to os.sep. Probably it's arguable solution, but at this
>>>> stage of development most of the code works well with os.sep internally
>>>> rather than '/'-only. (modified: bzrlib/branch.py)
>>> Yes, it's definitely arguable.  In fact, it's wrong for http branches
>>> and sftp branches right now.  This is why I didn't use os.path.basename.
>>>
>>> I thought we were going to use file URLs instead of paths anywhere where
>>> we were supporting URLs.  What's the plan here?
>> AIUI the plan is that:
>>  * The Transport layer takes URLs only.
>>  * Branches live at a URL
>>  * Trees take file paths (absolute or relative) only.
> 
> That sounds good to me.  The plan, as I understand it, is that we will
> also use / for paths internally, evne on win32.
> 
> Perhaps Branches should have base_url and base_path properties, where
> base_path raises an exception if it's not a local branch.
> 

I am planning on changing things so that all paths internally are
forward slashes, and just at the interface layer do we deal with
converting backwards slashes into forward ones.

John
=:->


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 249 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20051125/131edc8b/attachment.pgp