2005-05-21 17:40:01 +08:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2005 Junio C Hamano
|
2010-08-24 01:17:03 +08:00
|
|
|
* Copyright (C) 2010 Google Inc.
|
2005-05-21 17:40:01 +08:00
|
|
|
*/
|
|
|
|
#include "cache.h"
|
|
|
|
#include "diff.h"
|
|
|
|
#include "diffcore.h"
|
2010-08-24 01:17:03 +08:00
|
|
|
#include "xdiff-interface.h"
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
#include "kwset.h"
|
2010-08-24 01:17:03 +08:00
|
|
|
|
2011-10-07 00:50:28 +08:00
|
|
|
typedef int (*pickaxe_fn)(struct diff_filepair *p, struct diff_options *o, regex_t *regexp, kwset_t kws);
|
|
|
|
|
|
|
|
static void pickaxe(struct diff_queue_struct *q, struct diff_options *o,
|
|
|
|
regex_t *regexp, kwset_t kws, pickaxe_fn fn)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
struct diff_queue_struct outq;
|
|
|
|
|
|
|
|
DIFF_QUEUE_CLEAR(&outq);
|
|
|
|
|
|
|
|
if (o->pickaxe_opts & DIFF_PICKAXE_ALL) {
|
|
|
|
/* Showing the whole changeset if needle exists */
|
|
|
|
for (i = 0; i < q->nr; i++) {
|
|
|
|
struct diff_filepair *p = q->queue[i];
|
|
|
|
if (fn(p, o, regexp, kws))
|
|
|
|
return; /* do not munge the queue */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Otherwise we will clear the whole queue by copying
|
|
|
|
* the empty outq at the end of this function, but
|
|
|
|
* first clear the current entries in the queue.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < q->nr; i++)
|
|
|
|
diff_free_filepair(q->queue[i]);
|
|
|
|
} else {
|
|
|
|
/* Showing only the filepairs that has the needle */
|
|
|
|
for (i = 0; i < q->nr; i++) {
|
|
|
|
struct diff_filepair *p = q->queue[i];
|
|
|
|
if (fn(p, o, regexp, kws))
|
|
|
|
diff_q(&outq, p);
|
|
|
|
else
|
|
|
|
diff_free_filepair(p);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
free(q->queue);
|
|
|
|
*q = outq;
|
|
|
|
}
|
|
|
|
|
2010-08-24 01:17:03 +08:00
|
|
|
struct diffgrep_cb {
|
|
|
|
regex_t *regexp;
|
|
|
|
int hit;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void diffgrep_consume(void *priv, char *line, unsigned long len)
|
|
|
|
{
|
|
|
|
struct diffgrep_cb *data = priv;
|
|
|
|
regmatch_t regmatch;
|
|
|
|
int hold;
|
|
|
|
|
|
|
|
if (line[0] != '+' && line[0] != '-')
|
|
|
|
return;
|
|
|
|
if (data->hit)
|
|
|
|
/*
|
|
|
|
* NEEDSWORK: we should have a way to terminate the
|
|
|
|
* caller early.
|
|
|
|
*/
|
|
|
|
return;
|
|
|
|
/* Yuck -- line ought to be "const char *"! */
|
|
|
|
hold = line[len];
|
|
|
|
line[len] = '\0';
|
|
|
|
data->hit = !regexec(data->regexp, line + 1, 1, ®match, 0);
|
|
|
|
line[len] = hold;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void fill_one(struct diff_filespec *one,
|
|
|
|
mmfile_t *mf, struct userdiff_driver **textconv)
|
|
|
|
{
|
|
|
|
if (DIFF_FILE_VALID(one)) {
|
|
|
|
*textconv = get_textconv(one);
|
|
|
|
mf->size = fill_textconv(*textconv, one, &mf->ptr);
|
|
|
|
} else {
|
|
|
|
memset(mf, 0, sizeof(*mf));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-10-07 00:50:18 +08:00
|
|
|
static int diff_grep(struct diff_filepair *p, struct diff_options *o,
|
|
|
|
regex_t *regexp, kwset_t kws)
|
2010-08-24 01:17:03 +08:00
|
|
|
{
|
|
|
|
regmatch_t regmatch;
|
|
|
|
struct userdiff_driver *textconv_one = NULL;
|
|
|
|
struct userdiff_driver *textconv_two = NULL;
|
|
|
|
mmfile_t mf1, mf2;
|
|
|
|
int hit;
|
|
|
|
|
|
|
|
if (diff_unmodified_pair(p))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
fill_one(p->one, &mf1, &textconv_one);
|
|
|
|
fill_one(p->two, &mf2, &textconv_two);
|
|
|
|
|
|
|
|
if (!mf1.ptr) {
|
|
|
|
if (!mf2.ptr)
|
|
|
|
return 0; /* ignore unmerged */
|
|
|
|
/* created "two" -- does it have what we are looking for? */
|
diff_grep: use textconv buffers for add/deleted files
If you use "-G" to grep a diff, we will apply a configured
textconv filter to the data before generating the diff.
However, if the diff is an addition or deletion, we do not
bother running the diff at all, and just look for the token
in the added (or removed) content. This works because we
know that the diff must contain every line of content.
However, while we used the textconv-derived buffers in the
regular diff, we accidentally passed the original unmodified
buffers to regexec when checking the added or removed
content. This could lead to an incorrect answer.
Worse, in some cases we might have a textconv buffer but no
original buffer (e.g., if we pulled the textconv data from
cache, or if we reused a working tree file when generating
it). In that case, we could actually feed NULL to regexec
and segfault.
Reported-by: Peter Oberndorfer <kumbayo84@arcor.de>
Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 19:40:00 +08:00
|
|
|
hit = !regexec(regexp, mf2.ptr, 1, ®match, 0);
|
2010-08-24 01:17:03 +08:00
|
|
|
} else if (!mf2.ptr) {
|
|
|
|
/* removed "one" -- did it have what we are looking for? */
|
diff_grep: use textconv buffers for add/deleted files
If you use "-G" to grep a diff, we will apply a configured
textconv filter to the data before generating the diff.
However, if the diff is an addition or deletion, we do not
bother running the diff at all, and just look for the token
in the added (or removed) content. This works because we
know that the diff must contain every line of content.
However, while we used the textconv-derived buffers in the
regular diff, we accidentally passed the original unmodified
buffers to regexec when checking the added or removed
content. This could lead to an incorrect answer.
Worse, in some cases we might have a textconv buffer but no
original buffer (e.g., if we pulled the textconv data from
cache, or if we reused a working tree file when generating
it). In that case, we could actually feed NULL to regexec
and segfault.
Reported-by: Peter Oberndorfer <kumbayo84@arcor.de>
Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 19:40:00 +08:00
|
|
|
hit = !regexec(regexp, mf1.ptr, 1, ®match, 0);
|
2010-08-24 01:17:03 +08:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* We have both sides; need to run textual diff and see if
|
|
|
|
* the pattern appears on added/deleted lines.
|
|
|
|
*/
|
|
|
|
struct diffgrep_cb ecbdata;
|
|
|
|
xpparam_t xpp;
|
|
|
|
xdemitconf_t xecfg;
|
|
|
|
|
|
|
|
memset(&xpp, 0, sizeof(xpp));
|
|
|
|
memset(&xecfg, 0, sizeof(xecfg));
|
|
|
|
ecbdata.regexp = regexp;
|
|
|
|
ecbdata.hit = 0;
|
|
|
|
xecfg.ctxlen = o->context;
|
|
|
|
xecfg.interhunkctxlen = o->interhunkcontext;
|
|
|
|
xdi_diff_outf(&mf1, &mf2, diffgrep_consume, &ecbdata,
|
|
|
|
&xpp, &xecfg);
|
|
|
|
hit = ecbdata.hit;
|
|
|
|
}
|
|
|
|
if (textconv_one)
|
|
|
|
free(mf1.ptr);
|
|
|
|
if (textconv_two)
|
|
|
|
free(mf2.ptr);
|
|
|
|
return hit;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void diffcore_pickaxe_grep(struct diff_options *o)
|
|
|
|
{
|
2011-10-07 00:50:28 +08:00
|
|
|
int err;
|
2010-08-24 01:17:03 +08:00
|
|
|
regex_t regex;
|
2012-02-21 17:02:46 +08:00
|
|
|
int cflags = REG_EXTENDED | REG_NEWLINE;
|
2010-08-24 01:17:03 +08:00
|
|
|
|
2012-02-21 17:02:46 +08:00
|
|
|
if (DIFF_OPT_TST(o, PICKAXE_IGNORE_CASE))
|
|
|
|
cflags |= REG_ICASE;
|
|
|
|
|
|
|
|
err = regcomp(®ex, o->pickaxe, cflags);
|
2010-08-24 01:17:03 +08:00
|
|
|
if (err) {
|
|
|
|
char errbuf[1024];
|
|
|
|
regerror(err, ®ex, errbuf, 1024);
|
|
|
|
regfree(®ex);
|
|
|
|
die("invalid log-grep regex: %s", errbuf);
|
|
|
|
}
|
|
|
|
|
2011-10-07 00:50:28 +08:00
|
|
|
pickaxe(&diff_queued_diff, o, ®ex, NULL, diff_grep);
|
2010-08-24 01:17:03 +08:00
|
|
|
|
2011-10-07 00:14:55 +08:00
|
|
|
regfree(®ex);
|
2010-08-24 01:17:03 +08:00
|
|
|
return;
|
|
|
|
}
|
2005-05-21 17:40:01 +08:00
|
|
|
|
2012-10-28 20:27:12 +08:00
|
|
|
static unsigned int contains(mmfile_t *mf, struct diff_options *o,
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
regex_t *regexp, kwset_t kws)
|
2005-05-21 17:40:01 +08:00
|
|
|
{
|
2005-07-24 07:35:25 +08:00
|
|
|
unsigned int cnt;
|
2009-03-03 07:00:55 +08:00
|
|
|
unsigned long sz;
|
2005-05-21 17:40:01 +08:00
|
|
|
const char *data;
|
2005-07-24 07:35:25 +08:00
|
|
|
|
2012-10-28 20:27:12 +08:00
|
|
|
sz = mf->size;
|
|
|
|
data = mf->ptr;
|
2005-07-24 07:35:25 +08:00
|
|
|
cnt = 0;
|
|
|
|
|
2006-03-29 08:16:33 +08:00
|
|
|
if (regexp) {
|
|
|
|
regmatch_t regmatch;
|
|
|
|
int flags = 0;
|
|
|
|
|
2009-03-17 02:38:42 +08:00
|
|
|
assert(data[sz] == '\0');
|
2006-03-29 08:16:33 +08:00
|
|
|
while (*data && !regexec(regexp, data, 1, ®match, flags)) {
|
|
|
|
flags |= REG_NOTBOL;
|
2009-03-17 02:38:42 +08:00
|
|
|
data += regmatch.rm_eo;
|
|
|
|
if (*data && regmatch.rm_so == regmatch.rm_eo)
|
|
|
|
data++;
|
2005-07-24 07:35:25 +08:00
|
|
|
cnt++;
|
|
|
|
}
|
2006-03-29 08:16:33 +08:00
|
|
|
|
|
|
|
} else { /* Classic exact string match */
|
2009-03-03 07:00:55 +08:00
|
|
|
while (sz) {
|
2011-10-07 00:50:06 +08:00
|
|
|
struct kwsmatch kwsm;
|
|
|
|
size_t offset = kwsexec(kws, data, sz, &kwsm);
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
const char *found;
|
|
|
|
if (offset == -1)
|
2009-03-03 07:00:55 +08:00
|
|
|
break;
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
else
|
|
|
|
found = data + offset;
|
2011-10-07 00:50:06 +08:00
|
|
|
sz -= found - data + kwsm.size[0];
|
|
|
|
data = found + kwsm.size[0];
|
2009-03-03 07:00:55 +08:00
|
|
|
cnt++;
|
2006-03-29 08:16:33 +08:00
|
|
|
}
|
2005-07-24 07:35:25 +08:00
|
|
|
}
|
|
|
|
return cnt;
|
2005-05-21 17:40:01 +08:00
|
|
|
}
|
|
|
|
|
2011-10-07 00:50:06 +08:00
|
|
|
static int has_changes(struct diff_filepair *p, struct diff_options *o,
|
|
|
|
regex_t *regexp, kwset_t kws)
|
2011-10-07 00:26:24 +08:00
|
|
|
{
|
2012-10-28 20:27:12 +08:00
|
|
|
struct userdiff_driver *textconv_one = get_textconv(p->one);
|
|
|
|
struct userdiff_driver *textconv_two = get_textconv(p->two);
|
|
|
|
mmfile_t mf1, mf2;
|
|
|
|
int ret;
|
|
|
|
|
2012-10-28 20:34:06 +08:00
|
|
|
if (!o->pickaxe[0])
|
|
|
|
return 0;
|
|
|
|
|
2012-10-28 20:27:12 +08:00
|
|
|
/*
|
|
|
|
* If we have an unmodified pair, we know that the count will be the
|
|
|
|
* same and don't even have to load the blobs. Unless textconv is in
|
|
|
|
* play, _and_ we are using two different textconv filters (e.g.,
|
|
|
|
* because a pair is an exact rename with different textconv attributes
|
|
|
|
* for each side, which might generate different content).
|
|
|
|
*/
|
|
|
|
if (textconv_one == textconv_two && diff_unmodified_pair(p))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
fill_one(p->one, &mf1, &textconv_one);
|
|
|
|
fill_one(p->two, &mf2, &textconv_two);
|
|
|
|
|
|
|
|
if (!mf1.ptr) {
|
|
|
|
if (!mf2.ptr)
|
|
|
|
ret = 0; /* ignore unmerged */
|
2011-10-07 00:26:24 +08:00
|
|
|
/* created */
|
2012-10-28 20:27:12 +08:00
|
|
|
ret = contains(&mf2, o, regexp, kws) != 0;
|
2011-10-07 00:26:24 +08:00
|
|
|
}
|
2012-10-28 20:27:12 +08:00
|
|
|
else if (!mf2.ptr) /* removed */
|
|
|
|
ret = contains(&mf1, o, regexp, kws) != 0;
|
|
|
|
else
|
|
|
|
ret = contains(&mf1, o, regexp, kws) !=
|
|
|
|
contains(&mf2, o, regexp, kws);
|
|
|
|
|
|
|
|
if (textconv_one)
|
|
|
|
free(mf1.ptr);
|
|
|
|
if (textconv_two)
|
|
|
|
free(mf2.ptr);
|
|
|
|
diff_free_filespec_data(p->one);
|
|
|
|
diff_free_filespec_data(p->two);
|
|
|
|
|
|
|
|
return ret;
|
2011-10-07 00:26:24 +08:00
|
|
|
}
|
|
|
|
|
2010-08-24 01:17:03 +08:00
|
|
|
static void diffcore_pickaxe_count(struct diff_options *o)
|
2005-05-21 17:40:01 +08:00
|
|
|
{
|
2010-09-01 04:44:39 +08:00
|
|
|
const char *needle = o->pickaxe;
|
|
|
|
int opts = o->pickaxe_opts;
|
2005-05-21 17:40:01 +08:00
|
|
|
unsigned long len = strlen(needle);
|
2006-03-29 08:16:33 +08:00
|
|
|
regex_t regex, *regexp = NULL;
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
kwset_t kws = NULL;
|
2005-05-21 17:40:01 +08:00
|
|
|
|
2006-03-29 08:16:33 +08:00
|
|
|
if (opts & DIFF_PICKAXE_REGEX) {
|
|
|
|
int err;
|
|
|
|
err = regcomp(®ex, needle, REG_EXTENDED | REG_NEWLINE);
|
|
|
|
if (err) {
|
|
|
|
/* The POSIX.2 people are surely sick */
|
|
|
|
char errbuf[1024];
|
|
|
|
regerror(err, ®ex, errbuf, 1024);
|
|
|
|
regfree(®ex);
|
|
|
|
die("invalid pickaxe regex: %s", errbuf);
|
|
|
|
}
|
|
|
|
regexp = ®ex;
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
} else {
|
2012-02-21 17:02:46 +08:00
|
|
|
kws = kwsalloc(DIFF_OPT_TST(o, PICKAXE_IGNORE_CASE)
|
|
|
|
? tolower_trans_tbl : NULL);
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
kwsincr(kws, needle, len);
|
|
|
|
kwsprep(kws);
|
2006-03-29 08:16:33 +08:00
|
|
|
}
|
|
|
|
|
2011-10-07 00:50:28 +08:00
|
|
|
pickaxe(&diff_queued_diff, o, regexp, kws, has_changes);
|
2011-10-07 00:23:11 +08:00
|
|
|
|
2010-10-05 06:51:47 +08:00
|
|
|
if (opts & DIFF_PICKAXE_REGEX)
|
2006-03-29 08:16:33 +08:00
|
|
|
regfree(®ex);
|
Use kwset in pickaxe
Benchmarks in the hot cache case:
before:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
47,092,744 cache-misses # 2.825 M/sec ( +- 1.607% )
123,368,389 cache-references # 7.400 M/sec ( +- 0.812% )
330,040,998 branch-misses # 3.134 % ( +- 0.257% )
10,530,896,750 branches # 631.663 M/sec ( +- 0.121% )
62,037,201,030 instructions # 1.399 IPC ( +- 0.142% )
44,331,294,321 cycles # 2659.073 M/sec ( +- 0.326% )
96,794 page-faults # 0.006 M/sec ( +- 11.952% )
25 CPU-migrations # 0.000 M/sec ( +- 25.266% )
1,424 context-switches # 0.000 M/sec ( +- 0.540% )
16671.708650 task-clock-msecs # 0.997 CPUs ( +- 0.343% )
16.728692052 seconds time elapsed ( +- 0.344% )
after:
$ perf stat --repeat=5 git log -Sqwerty
Performance counter stats for 'git log -Sqwerty' (5 runs):
51,385,522 cache-misses # 4.619 M/sec ( +- 0.565% )
129,177,880 cache-references # 11.611 M/sec ( +- 0.219% )
319,222,775 branch-misses # 6.946 % ( +- 0.134% )
4,595,913,233 branches # 413.086 M/sec ( +- 0.112% )
31,395,042,533 instructions # 1.062 IPC ( +- 0.129% )
29,558,348,598 cycles # 2656.740 M/sec ( +- 0.204% )
93,224 page-faults # 0.008 M/sec ( +- 4.487% )
19 CPU-migrations # 0.000 M/sec ( +- 10.425% )
950 context-switches # 0.000 M/sec ( +- 0.360% )
11125.796039 task-clock-msecs # 0.997 CPUs ( +- 0.239% )
11.164216599 seconds time elapsed ( +- 0.240% )
So the kwset code is about 33% faster.
Signed-off-by: Fredrik Kuivinen <frekui@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-21 06:41:57 +08:00
|
|
|
else
|
|
|
|
kwsfree(kws);
|
2005-05-21 17:40:01 +08:00
|
|
|
return;
|
|
|
|
}
|
2010-08-24 01:17:03 +08:00
|
|
|
|
|
|
|
void diffcore_pickaxe(struct diff_options *o)
|
|
|
|
{
|
|
|
|
/* Might want to warn when both S and G are on; I don't care... */
|
|
|
|
if (o->pickaxe_opts & DIFF_PICKAXE_KIND_G)
|
2010-10-05 06:51:48 +08:00
|
|
|
diffcore_pickaxe_grep(o);
|
2010-08-24 01:17:03 +08:00
|
|
|
else
|
2010-10-05 06:51:48 +08:00
|
|
|
diffcore_pickaxe_count(o);
|
2010-08-24 01:17:03 +08:00
|
|
|
}
|