[PATCH 3 of 5 V2] ancestor: add a class to generate ancestors lazily

[Augie Fackler]

On Dec 14, 2012, at 11:26 PM, Kevin Bullock <kbullock+mercurial at ringworld.org> wrote:

> On 14 Dec 2012, at 7:25 PM, Siddharth Agarwal wrote:
> 
>> On 12/14/2012 04:13 PM, Siddharth Agarwal wrote:
>>> +class lazyancestorset(object):
>>> +    def __init__(self, repo, revs, increvs, stoprev=0):
>> 
>> I just realized Kevin asked me to pass in a changelog here, and I passed in a repo instead. Passing in repos seems to be a bit more common in general, but either way's fine with me.
>> 
>> The docstring refers to pfunc instead of repo -- I've fixed that locally.
> 
> I'd lean toward passing in the changelog directly, but I'll also defer to those who've been in this codebase longer than I have. Others have opinions?
> 

I think I'd lean that way too, but only slightly.