[PATCH] Use enhanced diff_delta() in the similarity estimator.

The diff_delta() interface was extended to reject generating too big a
delta while we were working on the packed GIT archive format.

Take advantage of that when generating delta in the similarity estimator
used in diffcore-rename.c

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/diffcore-rename.c b/diffcore-rename.c
index 29609c7..6a52699 100644 (file)
--- a/diffcore-rename.c
+++ b/diffcore-rename.c
@@ -136,6 +136,7 @@ static int estimate_similarity(struct diff_filespec *src,
         */
        void *delta;
        unsigned long delta_size, base_size, src_copied, literal_added;
+       unsigned long delta_limit;
        int score;
 
        /* We deal only with regular files.  Symlink renames are handled
@@ -163,9 +164,13 @@ static int estimate_similarity(struct diff_filespec *src,
        if (diff_populate_filespec(src, 0) || diff_populate_filespec(dst, 0))
                return 0; /* error but caught downstream */
 
+       delta_limit = base_size * (MAX_SCORE-minimum_score) / MAX_SCORE;
        delta = diff_delta(src->data, src->size,
                           dst->data, dst->size,
-                          &delta_size, ~0UL);
+                          &delta_size, delta_limit);
+       if (!delta)
+               /* If delta_limit is exceeded, we have too much differences */
+               return 0;
 
        /* A delta that has a lot of literal additions would have
         * big delta_size no matter what else it does.