[MERGE] Faster commit

[Robert Collins]

On Mon, 2007-06-04 at 13:28 -0500, John Arbash Meinel wrote:
> 
> 
> I'm okay with relaxing our requirements (for kind()), and not being backwards
> compatible. But we should be explicit about when we are trying to do what.
> 
> My basic feeling is that 'kind()' has always been a mostly internal function,
> so it is okay to break it a bit. But "id2path()" is used a lot more, so it
> should be deprecated properly.
> 
> Also, if we really want a double-compatible api, then all functions should be
> abstracted into this 2-layer (kind() just calls _kind()). So that people
> *implementing* the api are writing a different function than people *calling*
> the api.
> 
> I don't know that we really want that level of compatibility (especially when
> you consider the overhead for performance). But I did want to mention it.

I think that the harder a change is to accomodate the more we should
care about compatability.

These changes are easy to deal with, so as long as they are clearly
noted (we've done this style of change before, in 0.15), its ok :).

-Rob
-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20070605/99d9e841/attachment.pgp