Re: [PATCH] technical-docs: document tree-walking API

From: Stephen Boyd <bebarino@gmail.com>
Date: Aug 4, 2009, 06:00 UTC
Message-ID: <4A77CE9B.2060102@gmail.com>
In-Reply-To: <7v4oso87fy.fsf@alter.siamese.dyndns.org>

Junio C Hamano wrote:

Show 5 quoted lines

> I think the primary author is Linus for this API, but once you finished
> the documentation, I do not think "authors" section is particularly
> needed (git knows who the code came from reasonably well anyway).  The
> names of people in these stub documents were there for people like you;
> see 530e741 (Start preparing the API documents., 2007-11-24).

I'm ok with dropping the Authors section then.

Show 7 quoted lines

>
> It might be a bit misleading to say that the path member of struct
> name_entry is a "pathname".  It is _one_ path component, and by stringing
> multiple struct traverse_info's (each of which has one struct name_entry)
> together with their prev pointer, make_traverse_path() can reconstruct the
> full pathname from the toplevel down to the level you have traversed into.
> We may want to say "path component" to clarify the distinction.

This sounds reasonable for the traversing section. I'm a bit confused about how it would work for the walking section though. I don't get the notion of path components there because the API is focused around one level of a tree. Maybe it should just be called the entry's name in that case?