Re: [PATCH 03/19] reset.c: pass pathspec around instead of (prefix, argv) pair

On Fri, Jan 11, 2013 at 6:09 AM, Junio C Hamano [off-list ref] wrote:

quoted

Or I could hold off nd/parse-pathspec if this series has a better
chance of graduation first. Decision?

I am greedy and want to have both ;-)

Apparently I have no problems with your being greedy.

There is no textual conflict between the two topics at the moment,
but because the ultimate goal of your series is to remove all uses
of the pathspec.raw[] field outside the implementation of pathspec
matching, it might help to rename the field to _private_raw (or
remove it), and either make get_pathspec() private or disappear, to
ensure that the compiler will help us catching semantic conflicts
with new users of it at a late stage of your series.

There are still some uses for get_pathspec() and new call sites won't
cause big problems because they would need init_pathspec() to convert
get_pathspec() results to struct pathspec. I will rename raw[] though.
-- 
Duy