[brisbane-core MERGE] CHKInventory.iter_non_root_entries()

[Ian Clatworthy]

Aaron Bentley wrote:
> Ian Clatworthy wrote:
>> It seems that operations that walk directories in CHKInventories
>> are slow, e.g. directories() and iter_changes(). As it turns out
>> though, I don't *need* the path returned by iter_changes() in
>> fast-import - just the inventories entries while loading texts.
> 
> Okay, but why skip the root entry?  Aren't we trying to stop treating
> the root entry as special?

The current user of the API - loading texts during a commit - doesn't
want it so it does simplify the calling code not returning it. I don't
really mind either way though - I was mostly reimplementing the entries()
API without the path bit. My original cut at this gave entries() an
include-path=True parameter but that didn't turn out because the client
code wanted an iterator, not list.

I'm happy to go with whatever API name and behaviour you recommend.

Ian C.