[MERGE] Branch.iter_merge_sorted_revisions

[Ian Clatworthy]

Ian Clatworthy wrote:
> John Arbash Meinel wrote:

>> 1) I think we need to be thinking sooner rather than later about
>> handling partial ancestries.
> 
> Shall do next.

Done.

>> 2) As written now, this doesn't actually win us anything, because it
>> doesn't actually cache the merge_sorted results.
> 
> Maybe a separate patch?

The sorted graph is now cached. If we want to ditch the revid-to-revno
cache, can we do that as a separate patch?

>> 3) I would like to see us not include 'seq_num' as part of the generated
>> output. (Explicitly stripping it.) I don't think any code actually makes
>> use of it, and it is ill-defined when we are evaluating only partial
>> history.
> 
> I agree. If anything, it makes chaining of caches harder. The layer
> above can always use enumerate() if they need a sequence number.

Done.

Ian C.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: iter-merge-sorted-revisions-3.patch
Type: text/x-diff
Size: 18441 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20090124/608d613a/attachment-0001.bin