diffcore-rename: improve estimate_similarity() heuristics

The logic to quickly dismiss potential rename pairs was broken.  It
would too eagerly dismiss possible renames when all of the difference
was due to pure new data (or deleted data).

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

diffcore-rename.c

@@ -170,7 +170,7 @@ static int estimate_similarity(struct diff_filespec *src,
 	 * and the final score computation below would not have a
 	 * divide-by-zero issue.
 	 */
-	if (base_size * (MAX_SCORE-minimum_score) < delta_size * MAX_SCORE)
+	if (max_size * (MAX_SCORE-minimum_score) < delta_size * MAX_SCORE)
 		return 0;
 
 	if (!src->cnt_data && diff_populate_filespec(src, 0))