@@ -990,12 +990,19 @@ static void fill_origin_fingerprints(struct blame_origin *o, mmfile_t *file)
return;
o->num_lines = find_line_starts(&line_starts, o->file.ptr,
o->file.size);
- /* TODO: Will fill in fingerprints in a future commit */
+ o->fingerprints = xcalloc(sizeof(struct fingerprint), o->num_lines);
+ get_line_fingerprints(o->fingerprints, o->file.ptr, line_starts,
+ 0, o->num_lines);
free(line_starts);
}
static void drop_origin_fingerprints(struct blame_origin *o)
{
+ if (o->fingerprints) {
+ free_line_fingerprints(o->fingerprints, o->num_lines);
+ o->num_lines = 0;
+ FREE_AND_NULL(o->fingerprints);
+ }
}
/*
@@ -1573,9 +1580,34 @@ static int are_lines_adjacent(struct blame_line_tracker *first,
first->s_lno + 1 == second->s_lno;
}
+static int scan_parent_range(struct fingerprint *p_fps,
+ struct fingerprint *t_fps, int t_idx,
+ int from, int nr_lines)
+{
+ int sim, p_idx;
+ #define FINGERPRINT_FILE_THRESHOLD 10
+ int best_sim_val = FINGERPRINT_FILE_THRESHOLD;
+ int best_sim_idx = -1;
+
+ for (p_idx = from; p_idx < from + nr_lines; p_idx++) {
+ sim = fingerprint_similarity(&t_fps[t_idx], &p_fps[p_idx]);
+ if (sim < best_sim_val)
+ continue;
+ /* Break ties with the closest-to-target line number */
+ if (sim == best_sim_val && best_sim_idx != -1 &&
+ abs(best_sim_idx - t_idx) < abs(p_idx - t_idx))
+ continue;
+ best_sim_val = sim;
+ best_sim_idx = p_idx;
+ }
+ return best_sim_idx;
+}
+
/*
- * This cheap heuristic assigns lines in the chunk to their relative location in
- * the parent's chunk. Any additional lines are left with the target.
+ * The first pass checks the blame entry (from the target) against the parent's
+ * diff chunk. If that fails for a line, the second pass tries to match that
+ * line to any part of parent file. That catches cases where a change was
+ * broken into two chunks by 'context.'
*/
static void guess_line_blames(struct blame_origin *parent,
struct blame_origin *target,
@@ -1584,11 +1616,22 @@ static void guess_line_blames(struct blame_origin *parent,
{
int i, best_idx, target_idx;
int parent_slno = tlno + offset;
+ int *fuzzy_matches;
+ fuzzy_matches = fuzzy_find_matching_lines(parent, target,
+ tlno, parent_slno, same,
+ parent_len);
for (i = 0; i < same - tlno; i++) {
target_idx = tlno + i;
- best_idx = target_idx + offset;
- if (best_idx < parent_slno + parent_len) {
+ if (fuzzy_matches && fuzzy_matches[i] >= 0) {
+ best_idx = fuzzy_matches[i];
+ } else {
+ best_idx = scan_parent_range(parent->fingerprints,
+ target->fingerprints,
+ target_idx, 0,
+ parent->num_lines);
+ }
+ if (best_idx >= 0) {
line_blames[i].is_parent = 1;
line_blames[i].s_lno = best_idx;
} else {
@@ -1596,6 +1639,7 @@ static void guess_line_blames(struct blame_origin *parent,
line_blames[i].s_lno = target_idx;
}
}
+ free(fuzzy_matches);
}
/*
@@ -2372,6 +2416,12 @@ static void pass_blame(struct blame_scoreboard *sb, struct blame_origin *origin,
if (!porigin)
continue;
pass_blame_to_parent(sb, origin, porigin, 1);
+ /*
+ * Preemptively drop porigin so we can refresh the
+ * fingerprints if we use the parent again, which can
+ * occur if you ignore back-to-back commits.
+ */
+ drop_origin_blob(porigin);
if (!origin->suspects)
goto finish;
}
@@ -3,9 +3,6 @@
test_description='git blame ignore fuzzy heuristic'
. ./test-lib.sh
-# short circuit until blame has the fuzzy capabilities
-test_done
-
pick_author='s/^[0-9a-f^]* *(\([^ ]*\) .*/\1/'
# Each test is composed of 4 variables:
This commit integrates the fuzzy fingerprint heuristic into guess_line_blames(). We actually make two passes. The first pass uses the fuzzy algorithm to find a match within the current diff chunk. If that fails, the second pass searches the entire parent file for the best match. For an example of scanning the entire parent for a match, consider: commit-a 30) #include <sys/header_a.h> commit-b 31) #include <header_b.h> commit-c 32) #include <header_c.h> Then commit X alphabetizes them: commit-X 30) #include <header_b.h> commit-X 31) #include <header_c.h> commit-X 32) #include <sys/header_a.h> If we just check the parent's chunk (i.e. the first pass), we'd get: commit-b 30) #include <header_b.h> commit-c 31) #include <header_c.h> commit-X 32) #include <sys/header_a.h> That's because commit X actually consists of two chunks: one chunk is removing sys/header_a.h, then some context, and the second chunk is adding sys/header_a.h. If we scan the entire parent file, we get: commit-b 30) #include <header_b.h> commit-c 31) #include <header_c.h> commit-a 32) #include <sys/header_a.h> Signed-off-by: Barret Rhoden <brho@google.com> --- blame.c | 60 ++++++++++++++++++++++++++++++++--- t/t8014-blame-ignore-fuzzy.sh | 3 -- 2 files changed, 55 insertions(+), 8 deletions(-)