[MERGE] RevisionSpec.in_history() should raise nicer errors

[John Arbash Meinel]

Robert Collins wrote:

...

> Hi, I think this is appropriate for 0.11 and not 0.10. Its a step in the
> right direction, but there is more we can do - and doing that in 0.10
> would be way to much, as it is its already borderline IMO.
> 
> So I'd rather say 'lets get this really right for 0.11 and not change
> for 0.10'.
> 

since 0.11 opens up next week, attached is my updated diff.

It cleans up a lot of places that were directly instantiating a
RevisionSpec object (which used __new__ to return a child class), and
instead creates a factory function get_revision_spec(), and deprecates
directly instantiating a RevisionSpec object.


It merges RevisionSpec_int into RevisionSpec_revno, so that now we can
actually do:
bzr log -r 10:path/to/branch

Just like we could do
bzr log -r revno:10:path/to/branch

And it includes all the previous fixes and cleanups.

It uses a regex to look for integer style revno arguments. And gets rid
of the old support for 'bzr log -r10:20', since that has been deprecated
for a *really* long time (before we had explicit deprecation support)

John
=:->
-------------- next part --------------
A non-text attachment was scrubbed...
Name: revspec-55420.diff
Type: text/x-patch
Size: 49823 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060825/e8dc99e1/attachment.bin 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 254 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060825/e8dc99e1/attachment.pgp