git-commit-vandalism/blame.h
Derrick Stolee 0906ac2b54 blame: use changed-path Bloom filters
The changed-path Bloom filters help reduce the amount of tree
parsing required during history queries. Before calculating a
diff, we can ask the filter if a path changed between a commit
and its first parent. If the filter says "no" then we can move
on without parsing trees. If the filter says "maybe" then we
parse trees to discover if the answer is actually "yes" or "no".

When computing a blame, there is a section in find_origin() that
computes a diff between a commit and one of its parents. When this
is the first parent, we can check the Bloom filters before calling
diff_tree_oid().

In order to make this work with the blame machinery, we need to
initialize a struct bloom_key with the initial path. But also, we
need to add more keys to a list if a rename is detected. We then
check to see if _any_ of these keys answer "maybe" in the diff.

During development, I purposefully left out this "add a new key
when a rename is detected" to see if the test suite would catch
my error. That is how I discovered the issues with
GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS from the previous change.
With that change, we can feel some confidence in the coverage of
this change.

If a user requests copy detection using "git blame -C", then there
are more places where the set of "important" files can expand. I
do not know enough about how this happens in the blame machinery.
Thus, the Bloom filter integration is explicitly disabled in this
mode. A later change could expand the bloom_key data with an
appropriate call (or calls) to add_bloom_key().

If we did not disable this mode, then the following tests would
fail:

	t8003-blame-corner-cases.sh
	t8011-blame-split-file.sh

Generally, this is a performance enhancement and should not
change the behavior of 'git blame' in any way. If a repo has a
commit-graph file with computed changed-path Bloom filters, then
they should notice improved performance for their 'git blame'
commands.

Here are some example timings that I found by blaming some paths
in the Linux kernel repository:

 git blame arch/x86/kernel/topology.c >/dev/null

 Before: 0.83s
  After: 0.24s

 git blame kernel/time/time.c >/dev/null

 Before: 0.72s
  After: 0.24s

 git blame tools/perf/ui/stdio/hist.c >/dev/null

 Before: 0.27s
  After: 0.11s

I specifically looked for "deep" paths that were also edited many
times. As a counterpoint, the MAINTAINERS file was edited many
times but is located in the root tree. This means that the cost of
computing a diff relative to the pathspec is very small. Here are
the timings for that command:

 git blame MAINTAINERS >/dev/null

 Before: 20.1s
  After: 18.0s

These timings are the best of five. The worst-case runs were on the
order of 2.5 minutes for both cases. Note that the MAINTAINERS file
has 18,740 lines across 17,000+ commits. This happens to be one of
the cases where this change provides the least improvement.

The lack of improvement for the MAINTAINERS file and the relatively
modest improvement for the other examples can be easily explained.
The blame machinery needs to compute line-level diffs to determine
which lines were changed by each commit. That makes up a large
proportion of the computation time, and this change does not
attempt to improve on that section of the algorithm. The
MAINTAINERS file is large and changed often, so it takes time to
determine which lines were updated by which commit. In contrast,
the code files are much smaller, and it takes longer to comute
the line-by-line diff for a single patch on the Linux mailing
lists.

Outside of the "-C" integration, I believe there is little more to
gain from the changed-path Bloom filters for 'git blame' after this
patch.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-04-16 15:38:06 -07:00

197 lines
5.6 KiB
C

#ifndef BLAME_H
#define BLAME_H
#include "cache.h"
#include "commit.h"
#include "xdiff-interface.h"
#include "revision.h"
#include "prio-queue.h"
#include "diff.h"
#define PICKAXE_BLAME_MOVE 01
#define PICKAXE_BLAME_COPY 02
#define PICKAXE_BLAME_COPY_HARDER 04
#define PICKAXE_BLAME_COPY_HARDEST 010
#define BLAME_DEFAULT_MOVE_SCORE 20
#define BLAME_DEFAULT_COPY_SCORE 40
struct fingerprint;
/*
* One blob in a commit that is being suspected
*/
struct blame_origin {
int refcnt;
/* Record preceding blame record for this blob */
struct blame_origin *previous;
/* origins are put in a list linked via `next' hanging off the
* corresponding commit's util field in order to make finding
* them fast. The presence in this chain does not count
* towards the origin's reference count. It is tempting to
* let it count as long as the commit is pending examination,
* but even under circumstances where the commit will be
* present multiple times in the priority queue of unexamined
* commits, processing the first instance will not leave any
* work requiring the origin data for the second instance. An
* interspersed commit changing that would have to be
* preexisting with a different ancestry and with the same
* commit date in order to wedge itself between two instances
* of the same commit in the priority queue _and_ produce
* blame entries relevant for it. While we don't want to let
* us get tripped up by this case, it certainly does not seem
* worth optimizing for.
*/
struct blame_origin *next;
struct commit *commit;
/* `suspects' contains blame entries that may be attributed to
* this origin's commit or to parent commits. When a commit
* is being processed, all suspects will be moved, either by
* assigning them to an origin in a different commit, or by
* shipping them to the scoreboard's ent list because they
* cannot be attributed to a different commit.
*/
struct blame_entry *suspects;
mmfile_t file;
int num_lines;
struct fingerprint *fingerprints;
struct object_id blob_oid;
unsigned short mode;
/* guilty gets set when shipping any suspects to the final
* blame list instead of other commits
*/
char guilty;
char path[FLEX_ARRAY];
};
/*
* Each group of lines is described by a blame_entry; it can be split
* as we pass blame to the parents. They are arranged in linked lists
* kept as `suspects' of some unprocessed origin, or entered (when the
* blame origin has been finalized) into the scoreboard structure.
* While the scoreboard structure is only sorted at the end of
* processing (according to final image line number), the lists
* attached to an origin are sorted by the target line number.
*/
struct blame_entry {
struct blame_entry *next;
/* the first line of this group in the final image;
* internally all line numbers are 0 based.
*/
int lno;
/* how many lines this group has */
int num_lines;
/* the commit that introduced this group into the final image */
struct blame_origin *suspect;
/* the line number of the first line of this group in the
* suspect's file; internally all line numbers are 0 based.
*/
int s_lno;
/* how significant this entry is -- cached to avoid
* scanning the lines over and over.
*/
unsigned score;
int ignored;
int unblamable;
};
struct blame_bloom_data;
/*
* The current state of the blame assignment.
*/
struct blame_scoreboard {
/* the final commit (i.e. where we started digging from) */
struct commit *final;
/* Priority queue for commits with unassigned blame records */
struct prio_queue commits;
struct repository *repo;
struct rev_info *revs;
const char *path;
/*
* The contents in the final image.
* Used by many functions to obtain contents of the nth line,
* indexed with scoreboard.lineno[blame_entry.lno].
*/
const char *final_buf;
unsigned long final_buf_size;
/* linked list of blames */
struct blame_entry *ent;
struct oidset ignore_list;
/* look-up a line in the final buffer */
int num_lines;
int *lineno;
/* stats */
int num_read_blob;
int num_get_patch;
int num_commits;
/*
* blame for a blame_entry with score lower than these thresholds
* is not passed to the parent using move/copy logic.
*/
unsigned move_score;
unsigned copy_score;
/* use this file's contents as the final image */
const char *contents_from;
/* flags */
int reverse;
int show_root;
int xdl_opts;
int no_whole_file_rename;
int debug;
/* callbacks */
void(*on_sanity_fail)(struct blame_scoreboard *, int);
void(*found_guilty_entry)(struct blame_entry *, void *);
void *found_guilty_entry_data;
struct blame_bloom_data *bloom_data;
};
/*
* Origin is refcounted and usually we keep the blob contents to be
* reused.
*/
static inline struct blame_origin *blame_origin_incref(struct blame_origin *o)
{
if (o)
o->refcnt++;
return o;
}
void blame_origin_decref(struct blame_origin *o);
void blame_coalesce(struct blame_scoreboard *sb);
void blame_sort_final(struct blame_scoreboard *sb);
unsigned blame_entry_score(struct blame_scoreboard *sb, struct blame_entry *e);
void assign_blame(struct blame_scoreboard *sb, int opt);
const char *blame_nth_line(struct blame_scoreboard *sb, long lno);
void init_scoreboard(struct blame_scoreboard *sb);
void setup_scoreboard(struct blame_scoreboard *sb,
const char *path,
struct blame_origin **orig);
void setup_blame_bloom_data(struct blame_scoreboard *sb,
const char *path);
void cleanup_scoreboard(struct blame_scoreboard *sb);
struct blame_entry *blame_entry_prepend(struct blame_entry *head,
long start, long end,
struct blame_origin *o);
struct blame_origin *get_blame_suspects(struct commit *commit);
#endif /* BLAME_H */