rfc: remove "revision specs"

[Robert Collins]

On Tue, 2005-08-30 at 09:34 +0200, Matthieu Moy wrote:

> I'd like revision identifiers to be regexp-matchable in bzr too,
> although I'm not sure how it's possible to do it.
> bzr:///home/foo/bar/,rev=1040  could be nice, but OTOH,
> bzr://http://www.site.com/my/bzr/branch/,rev=1040  looks really ughly.

Couple of syntax nits here:
.../bar/,rev=1040 is different to .../bar,rev=1040 is.
The former is a parameter to an empty path element, the latter is a
parameter to the bar element. I'm suggesting we allow parameters to the
element that is the root of the branch.

bzr:///home/foo/bar <- if by that you mean a bzr branch on the local
file system, the relevant url scheme is 'file:' ->
'file:///home/foo/bar' unless we are defining a new access protocol.
(We're not at this point). One could defined bzr+file: if we want to be
able to define it but I think that that isn't useful at this point.

bzr://http://.. isn't valid std66 format for a url unless the entire
scheme is opaque - and even then, it would be badly escaped. Similarly
to your bzr:///home example, the right way to write it would be
bzr+http://www.site.com/my/bzr/branch,rev=1040 (If we extend http)
or
http://www.site.com/my/bzr/branch,rev=1040 (if we use webdav/normal
http)

> In brief, revision specifiers within a path are a necessary condition
> to have easily recognizable bzr revision specifiers, but it's not
> sufficient. Or perhaps forget about this all if this is too complex
> for too little benefit ...

Does the simpler way I'm writing them influence this feeling ?

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/20050831/43c1d498/attachment.pgp