Re: [PATCH 1/5] Add 'df_name_compare()' helper function

[David Kastrup <dak@gnu.org>]

Linus Torvalds <torvalds@linux-foundation.org> writes:

> This new helper is identical to base_name_compare(), except it compares
> conflicting directory/file entries as equal in order to help handling DF
> conflicts (thus the name).
>
> Note that while a directory name compares as equal to a regular file
> with the new helper, they then individually compare _differently_ to a
> filename that has a dot after the basename (because '\0' < '.' < '/').

Uh, that screams just bloody murder at me with regard to working on
sorted material.  You'll never even get into the situation where the
conflicting "equal" entries will be compared if you presorted them with
something in between.  Consider the case of a merge of

A:
blubb
blubb.hi

B:
blubb.hi
blubb/

Any traversal that is done reasonably efficiently will never compare
blubb to blubb/ and I don't see how to get around this.

Basically, I think you need a special traversal routine.  This routine
will push all non-directory entries during a parallel traverse into a
might-conflict-queue, appending a slash in the course.  The front of
this queue then gets compared with the next processed entry.  If it is
larger, it is kept at the front, if it is equal, the conflict gets
treated/resolved, if it is smaller, it gets popped (since it can't
conflict anymore).

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum