2005-06-26 05:42:43 +08:00
|
|
|
#include "cache.h"
|
|
|
|
#include "object.h"
|
|
|
|
#include "delta.h"
|
2005-06-29 05:21:02 +08:00
|
|
|
#include "pack.h"
|
2005-06-27 11:27:56 +08:00
|
|
|
#include "csum-file.h"
|
2006-02-12 09:54:18 +08:00
|
|
|
#include <sys/time.h>
|
2005-06-26 05:42:43 +08:00
|
|
|
|
2006-02-13 05:01:54 +08:00
|
|
|
static const char pack_usage[] = "git-pack-objects [-q] [--non-empty] [--local] [--incremental] [--window=N] [--depth=N] {--stdout | base-name} < object-list";
|
2005-06-26 05:42:43 +08:00
|
|
|
|
|
|
|
struct object_entry {
|
|
|
|
unsigned char sha1[20];
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
unsigned long size; /* uncompressed size */
|
|
|
|
unsigned long offset; /* offset into the final pack file (nonzero if already written) */
|
|
|
|
unsigned int depth; /* delta depth */
|
|
|
|
unsigned int hash; /* name hint hash */
|
2005-06-29 05:21:02 +08:00
|
|
|
enum object_type type;
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
unsigned long delta_size; /* delta data size (uncompressed) */
|
|
|
|
struct object_entry *delta; /* delta base object */
|
|
|
|
struct packed_git *in_pack; /* already in pack */
|
|
|
|
enum object_type in_pack_type; /* could be delta */
|
|
|
|
unsigned int in_pack_offset;
|
2005-06-26 05:42:43 +08:00
|
|
|
};
|
|
|
|
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
/*
|
|
|
|
* Objects we are going to pack are colected in objects array (dynamically
|
|
|
|
* expanded). nr_objects & nr_alloc controls this array. They are stored
|
|
|
|
* in the order we see -- typically rev-list --objects order that gives us
|
|
|
|
* nice "minimum seek" order.
|
|
|
|
*
|
|
|
|
* sorted-by-sha ans sorted-by-type are arrays of pointers that point at
|
|
|
|
* elements in the objects array. The former is used to build the pack
|
|
|
|
* index (lists object names in the ascending order to help offset lookup),
|
|
|
|
* and the latter is used to group similar things together by try_delta()
|
|
|
|
* heuristics.
|
|
|
|
*/
|
|
|
|
|
2005-07-04 06:34:04 +08:00
|
|
|
static unsigned char object_list_sha1[20];
|
2005-07-04 04:36:58 +08:00
|
|
|
static int non_empty = 0;
|
2005-10-14 06:38:28 +08:00
|
|
|
static int local = 0;
|
2005-07-04 04:08:40 +08:00
|
|
|
static int incremental = 0;
|
2005-06-26 05:42:43 +08:00
|
|
|
static struct object_entry **sorted_by_sha, **sorted_by_type;
|
|
|
|
static struct object_entry *objects = NULL;
|
|
|
|
static int nr_objects = 0, nr_alloc = 0;
|
|
|
|
static const char *base_name;
|
2005-06-27 13:01:46 +08:00
|
|
|
static unsigned char pack_file_sha1[20];
|
2006-02-13 05:01:54 +08:00
|
|
|
static int progress = 1;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
/*
|
|
|
|
* The object names in objects array are hashed with this hashtable,
|
|
|
|
* to help looking up the entry by object name. Binary search from
|
|
|
|
* sorted_by_sha is also possible but this was easier to code and faster.
|
|
|
|
* This hashtable is built after all the objects are seen.
|
|
|
|
*/
|
|
|
|
static int *object_ix = NULL;
|
|
|
|
static int object_ix_hashsz = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pack index for existing packs give us easy access to the offsets into
|
|
|
|
* corresponding pack file where each object's data starts, but the entries
|
|
|
|
* do not store the size of the compressed representation (uncompressed
|
|
|
|
* size is easily available by examining the pack entry header). We build
|
|
|
|
* a hashtable of existing packs (pack_revindex), and keep reverse index
|
|
|
|
* here -- pack index file is sorted by object name mapping to offset; this
|
|
|
|
* pack_revindex[].revindex array is an ordered list of offsets, so if you
|
|
|
|
* know the offset of an object, next offset is where its packed
|
|
|
|
* representation ends.
|
|
|
|
*/
|
|
|
|
struct pack_revindex {
|
|
|
|
struct packed_git *p;
|
|
|
|
unsigned long *revindex;
|
|
|
|
} *pack_revindex = NULL;
|
|
|
|
static int pack_revindex_hashsz = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* stats
|
|
|
|
*/
|
|
|
|
static int written = 0;
|
|
|
|
static int reused = 0;
|
|
|
|
|
|
|
|
static int pack_revindex_ix(struct packed_git *p)
|
|
|
|
{
|
|
|
|
unsigned int ui = (unsigned int) p;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
ui = ui ^ (ui >> 16); /* defeat structure alignment */
|
|
|
|
i = (int)(ui % pack_revindex_hashsz);
|
|
|
|
while (pack_revindex[i].p) {
|
|
|
|
if (pack_revindex[i].p == p)
|
|
|
|
return i;
|
|
|
|
if (++i == pack_revindex_hashsz)
|
|
|
|
i = 0;
|
|
|
|
}
|
|
|
|
return -1 - i;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prepare_pack_ix(void)
|
|
|
|
{
|
|
|
|
int num;
|
|
|
|
struct packed_git *p;
|
|
|
|
for (num = 0, p = packed_git; p; p = p->next)
|
|
|
|
num++;
|
|
|
|
if (!num)
|
|
|
|
return;
|
|
|
|
pack_revindex_hashsz = num * 11;
|
|
|
|
pack_revindex = xcalloc(sizeof(*pack_revindex), pack_revindex_hashsz);
|
|
|
|
for (p = packed_git; p; p = p->next) {
|
|
|
|
num = pack_revindex_ix(p);
|
|
|
|
num = - 1 - num;
|
|
|
|
pack_revindex[num].p = p;
|
|
|
|
}
|
|
|
|
/* revindex elements are lazily initialized */
|
|
|
|
}
|
|
|
|
|
|
|
|
static int cmp_offset(const void *a_, const void *b_)
|
|
|
|
{
|
|
|
|
unsigned long a = *(unsigned long *) a_;
|
|
|
|
unsigned long b = *(unsigned long *) b_;
|
|
|
|
if (a < b)
|
|
|
|
return -1;
|
|
|
|
else if (a == b)
|
|
|
|
return 0;
|
|
|
|
else
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Ordered list of offsets of objects in the pack.
|
|
|
|
*/
|
|
|
|
static void prepare_pack_revindex(struct pack_revindex *rix)
|
|
|
|
{
|
|
|
|
struct packed_git *p = rix->p;
|
|
|
|
int num_ent = num_packed_objects(p);
|
|
|
|
int i;
|
|
|
|
void *index = p->index_base + 256;
|
|
|
|
|
|
|
|
rix->revindex = xmalloc(sizeof(unsigned long) * (num_ent + 1));
|
|
|
|
for (i = 0; i < num_ent; i++) {
|
|
|
|
long hl = *((long *)(index + 24 * i));
|
|
|
|
rix->revindex[i] = ntohl(hl);
|
|
|
|
}
|
|
|
|
/* This knows the pack format -- the 20-byte trailer
|
|
|
|
* follows immediately after the last object data.
|
|
|
|
*/
|
|
|
|
rix->revindex[num_ent] = p->pack_size - 20;
|
|
|
|
qsort(rix->revindex, num_ent, sizeof(unsigned long), cmp_offset);
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long find_packed_object_size(struct packed_git *p,
|
|
|
|
unsigned long ofs)
|
|
|
|
{
|
|
|
|
int num;
|
|
|
|
int lo, hi;
|
|
|
|
struct pack_revindex *rix;
|
|
|
|
unsigned long *revindex;
|
|
|
|
num = pack_revindex_ix(p);
|
|
|
|
if (num < 0)
|
|
|
|
die("internal error: pack revindex uninitialized");
|
|
|
|
rix = &pack_revindex[num];
|
|
|
|
if (!rix->revindex)
|
|
|
|
prepare_pack_revindex(rix);
|
|
|
|
revindex = rix->revindex;
|
|
|
|
lo = 0;
|
|
|
|
hi = num_packed_objects(p) + 1;
|
|
|
|
do {
|
|
|
|
int mi = (lo + hi) / 2;
|
|
|
|
if (revindex[mi] == ofs) {
|
|
|
|
return revindex[mi+1] - ofs;
|
|
|
|
}
|
|
|
|
else if (ofs < revindex[mi])
|
|
|
|
hi = mi;
|
|
|
|
else
|
|
|
|
lo = mi + 1;
|
|
|
|
} while (lo < hi);
|
|
|
|
die("internal error: pack revindex corrupt");
|
|
|
|
}
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
static void *delta_against(void *buf, unsigned long size, struct object_entry *entry)
|
|
|
|
{
|
|
|
|
unsigned long othersize, delta_size;
|
|
|
|
char type[10];
|
|
|
|
void *otherbuf = read_sha1_file(entry->delta->sha1, type, &othersize);
|
|
|
|
void *delta_buf;
|
|
|
|
|
|
|
|
if (!otherbuf)
|
|
|
|
die("unable to read %s", sha1_to_hex(entry->delta->sha1));
|
2005-06-26 19:29:18 +08:00
|
|
|
delta_buf = diff_delta(otherbuf, othersize,
|
2005-06-29 14:49:56 +08:00
|
|
|
buf, size, &delta_size, 0);
|
2005-06-26 10:30:20 +08:00
|
|
|
if (!delta_buf || delta_size != entry->delta_size)
|
2005-06-26 05:42:43 +08:00
|
|
|
die("delta size changed");
|
|
|
|
free(buf);
|
|
|
|
free(otherbuf);
|
|
|
|
return delta_buf;
|
|
|
|
}
|
|
|
|
|
2005-06-29 05:21:02 +08:00
|
|
|
/*
|
|
|
|
* The per-object header is a pretty dense thing, which is
|
|
|
|
* - first byte: low four bits are "size", then three bits of "type",
|
|
|
|
* and the high bit is "size continues".
|
|
|
|
* - each byte afterwards: low seven bits are size continuation,
|
|
|
|
* with the high bit being "size continues"
|
|
|
|
*/
|
|
|
|
static int encode_header(enum object_type type, unsigned long size, unsigned char *hdr)
|
|
|
|
{
|
2005-06-29 13:15:57 +08:00
|
|
|
int n = 1;
|
2005-06-29 05:21:02 +08:00
|
|
|
unsigned char c;
|
|
|
|
|
|
|
|
if (type < OBJ_COMMIT || type > OBJ_DELTA)
|
|
|
|
die("bad type %d", type);
|
|
|
|
|
2005-06-29 13:15:57 +08:00
|
|
|
c = (type << 4) | (size & 15);
|
|
|
|
size >>= 4;
|
|
|
|
while (size) {
|
2005-06-29 05:21:02 +08:00
|
|
|
*hdr++ = c | 0x80;
|
2005-06-29 13:15:57 +08:00
|
|
|
c = size & 0x7f;
|
|
|
|
size >>= 7;
|
|
|
|
n++;
|
2005-06-29 05:21:02 +08:00
|
|
|
}
|
|
|
|
*hdr = c;
|
|
|
|
return n;
|
|
|
|
}
|
|
|
|
|
2005-06-27 11:27:56 +08:00
|
|
|
static unsigned long write_object(struct sha1file *f, struct object_entry *entry)
|
2005-06-26 05:42:43 +08:00
|
|
|
{
|
|
|
|
unsigned long size;
|
|
|
|
char type[10];
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
void *buf;
|
2005-06-29 05:21:02 +08:00
|
|
|
unsigned char header[10];
|
2005-06-26 05:42:43 +08:00
|
|
|
unsigned hdrlen, datalen;
|
2005-06-29 05:21:02 +08:00
|
|
|
enum object_type obj_type;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
2005-06-29 05:21:02 +08:00
|
|
|
obj_type = entry->type;
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
if (!entry->in_pack ||
|
|
|
|
(obj_type != entry->in_pack_type)) {
|
|
|
|
buf = read_sha1_file(entry->sha1, type, &size);
|
|
|
|
if (!buf)
|
|
|
|
die("unable to read %s", sha1_to_hex(entry->sha1));
|
|
|
|
if (size != entry->size)
|
|
|
|
die("object %s size inconsistency (%lu vs %lu)",
|
|
|
|
sha1_to_hex(entry->sha1), size, entry->size);
|
|
|
|
if (entry->delta) {
|
|
|
|
buf = delta_against(buf, size, entry);
|
|
|
|
size = entry->delta_size;
|
|
|
|
obj_type = OBJ_DELTA;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* The object header is a byte of 'type' followed by zero or
|
|
|
|
* more bytes of length. For deltas, the 20 bytes of delta
|
|
|
|
* sha1 follows that.
|
|
|
|
*/
|
|
|
|
hdrlen = encode_header(obj_type, size, header);
|
|
|
|
sha1write(f, header, hdrlen);
|
|
|
|
|
|
|
|
if (entry->delta) {
|
|
|
|
sha1write(f, entry->delta, 20);
|
|
|
|
hdrlen += 20;
|
|
|
|
}
|
|
|
|
datalen = sha1write_compressed(f, buf, size);
|
|
|
|
free(buf);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
else {
|
|
|
|
struct packed_git *p = entry->in_pack;
|
|
|
|
use_packed_git(p);
|
|
|
|
|
|
|
|
datalen = find_packed_object_size(p, entry->in_pack_offset);
|
|
|
|
buf = p->pack_base + entry->in_pack_offset;
|
|
|
|
sha1write(f, buf, datalen);
|
|
|
|
unuse_packed_git(p);
|
|
|
|
hdrlen = 0; /* not really */
|
|
|
|
reused++;
|
2005-06-29 05:21:02 +08:00
|
|
|
}
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
written++;
|
2005-06-26 05:42:43 +08:00
|
|
|
return hdrlen + datalen;
|
|
|
|
}
|
|
|
|
|
2005-06-29 08:49:27 +08:00
|
|
|
static unsigned long write_one(struct sha1file *f,
|
|
|
|
struct object_entry *e,
|
|
|
|
unsigned long offset)
|
|
|
|
{
|
|
|
|
if (e->offset)
|
|
|
|
/* offset starts from header size and cannot be zero
|
|
|
|
* if it is written already.
|
|
|
|
*/
|
|
|
|
return offset;
|
|
|
|
e->offset = offset;
|
|
|
|
offset += write_object(f, e);
|
2005-12-29 17:30:08 +08:00
|
|
|
/* if we are deltified, write out its base object. */
|
2005-06-29 08:49:27 +08:00
|
|
|
if (e->delta)
|
|
|
|
offset = write_one(f, e->delta, offset);
|
|
|
|
return offset;
|
|
|
|
}
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
static void write_pack_file(void)
|
|
|
|
{
|
|
|
|
int i;
|
2005-06-29 02:10:48 +08:00
|
|
|
struct sha1file *f;
|
2005-06-29 05:21:02 +08:00
|
|
|
unsigned long offset;
|
2005-06-26 05:42:43 +08:00
|
|
|
unsigned long mb;
|
2005-06-29 05:21:02 +08:00
|
|
|
struct pack_header hdr;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
2005-06-29 02:10:48 +08:00
|
|
|
if (!base_name)
|
|
|
|
f = sha1fd(1, "<stdout>");
|
|
|
|
else
|
2005-07-04 06:34:04 +08:00
|
|
|
f = sha1create("%s-%s.%s", base_name, sha1_to_hex(object_list_sha1), "pack");
|
2005-06-29 05:21:02 +08:00
|
|
|
hdr.hdr_signature = htonl(PACK_SIGNATURE);
|
2005-06-29 13:15:57 +08:00
|
|
|
hdr.hdr_version = htonl(PACK_VERSION);
|
2005-06-29 05:21:02 +08:00
|
|
|
hdr.hdr_entries = htonl(nr_objects);
|
|
|
|
sha1write(f, &hdr, sizeof(hdr));
|
|
|
|
offset = sizeof(hdr);
|
2005-06-29 08:49:27 +08:00
|
|
|
for (i = 0; i < nr_objects; i++)
|
|
|
|
offset = write_one(f, objects + i, offset);
|
|
|
|
|
2005-06-27 13:01:46 +08:00
|
|
|
sha1close(f, pack_file_sha1, 1);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void write_index_file(void)
|
|
|
|
{
|
|
|
|
int i;
|
2005-07-04 06:34:04 +08:00
|
|
|
struct sha1file *f = sha1create("%s-%s.%s", base_name, sha1_to_hex(object_list_sha1), "idx");
|
2005-06-26 05:42:43 +08:00
|
|
|
struct object_entry **list = sorted_by_sha;
|
|
|
|
struct object_entry **last = list + nr_objects;
|
|
|
|
unsigned int array[256];
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Write the first-level table (the list is sorted,
|
|
|
|
* but we use a 256-entry lookup to be able to avoid
|
2005-06-27 13:01:46 +08:00
|
|
|
* having to do eight extra binary search iterations).
|
2005-06-26 05:42:43 +08:00
|
|
|
*/
|
|
|
|
for (i = 0; i < 256; i++) {
|
|
|
|
struct object_entry **next = list;
|
|
|
|
while (next < last) {
|
|
|
|
struct object_entry *entry = *next;
|
|
|
|
if (entry->sha1[0] != i)
|
|
|
|
break;
|
|
|
|
next++;
|
|
|
|
}
|
|
|
|
array[i] = htonl(next - sorted_by_sha);
|
|
|
|
list = next;
|
|
|
|
}
|
2005-06-27 11:27:56 +08:00
|
|
|
sha1write(f, array, 256 * sizeof(int));
|
2005-06-26 05:42:43 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Write the actual SHA1 entries..
|
|
|
|
*/
|
|
|
|
list = sorted_by_sha;
|
2005-06-26 06:24:30 +08:00
|
|
|
for (i = 0; i < nr_objects; i++) {
|
2005-06-26 05:42:43 +08:00
|
|
|
struct object_entry *entry = *list++;
|
|
|
|
unsigned int offset = htonl(entry->offset);
|
2005-06-27 11:27:56 +08:00
|
|
|
sha1write(f, &offset, 4);
|
|
|
|
sha1write(f, entry->sha1, 20);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
2005-06-27 13:01:46 +08:00
|
|
|
sha1write(f, pack_file_sha1, 20);
|
|
|
|
sha1close(f, NULL, 1);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
2005-07-04 06:34:04 +08:00
|
|
|
static int add_object_entry(unsigned char *sha1, unsigned int hash)
|
2005-06-26 05:42:43 +08:00
|
|
|
{
|
|
|
|
unsigned int idx = nr_objects;
|
|
|
|
struct object_entry *entry;
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
struct packed_git *p;
|
|
|
|
unsigned int found_offset;
|
|
|
|
struct packed_git *found_pack;
|
|
|
|
|
|
|
|
found_pack = NULL;
|
|
|
|
for (p = packed_git; p; p = p->next) {
|
|
|
|
struct pack_entry e;
|
|
|
|
if (find_pack_entry_one(sha1, &e, p)) {
|
|
|
|
if (incremental)
|
|
|
|
return 0;
|
|
|
|
if (local && !p->pack_local)
|
|
|
|
return 0;
|
|
|
|
if (!found_pack) {
|
|
|
|
found_offset = e.offset;
|
|
|
|
found_pack = e.p;
|
2005-10-14 06:38:28 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2005-07-04 04:08:40 +08:00
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
if (idx >= nr_alloc) {
|
|
|
|
unsigned int needed = (idx + 1024) * 3 / 2;
|
|
|
|
objects = xrealloc(objects, needed * sizeof(*entry));
|
|
|
|
nr_alloc = needed;
|
|
|
|
}
|
|
|
|
entry = objects + idx;
|
|
|
|
memset(entry, 0, sizeof(*entry));
|
|
|
|
memcpy(entry->sha1, sha1, 20);
|
2005-06-27 06:27:28 +08:00
|
|
|
entry->hash = hash;
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
if (found_pack) {
|
|
|
|
entry->in_pack = found_pack;
|
|
|
|
entry->in_pack_offset = found_offset;
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
nr_objects = idx+1;
|
2005-07-04 06:34:04 +08:00
|
|
|
return 1;
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
static int locate_object_entry_hash(unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
unsigned int ui;
|
|
|
|
memcpy(&ui, sha1, sizeof(unsigned int));
|
|
|
|
i = ui % object_ix_hashsz;
|
|
|
|
while (0 < object_ix[i]) {
|
|
|
|
if (!memcmp(sha1, objects[object_ix[i]-1].sha1, 20))
|
|
|
|
return i;
|
|
|
|
if (++i == object_ix_hashsz)
|
|
|
|
i = 0;
|
|
|
|
}
|
|
|
|
return -1 - i;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct object_entry *locate_object_entry(unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int i = locate_object_entry_hash(sha1);
|
|
|
|
if (0 <= i)
|
|
|
|
return &objects[object_ix[i]-1];
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
static void check_object(struct object_entry *entry)
|
|
|
|
{
|
2005-06-27 18:34:06 +08:00
|
|
|
char type[20];
|
|
|
|
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
if (entry->in_pack) {
|
|
|
|
/* Check if it is delta, and the base is also an object
|
|
|
|
* we are going to pack. If so we will reuse the existing
|
|
|
|
* delta.
|
|
|
|
*/
|
|
|
|
unsigned char base[20];
|
|
|
|
unsigned long size;
|
|
|
|
struct object_entry *base_entry;
|
|
|
|
if (!check_reuse_pack_delta(entry->in_pack,
|
|
|
|
entry->in_pack_offset,
|
|
|
|
base, &size,
|
|
|
|
&entry->in_pack_type) &&
|
|
|
|
(base_entry = locate_object_entry(base))) {
|
|
|
|
/* We do not know depth at this point, but it
|
|
|
|
* does not matter. Getting delta_chain_length
|
|
|
|
* with packed_object_info_detail() is not so
|
|
|
|
* expensive, so we could do that later if we
|
|
|
|
* wanted to. Calling sha1_object_info to get
|
|
|
|
* the true size (and later an uncompressed
|
|
|
|
* representation) of deeply deltified object
|
|
|
|
* is quite expensive.
|
|
|
|
*/
|
|
|
|
entry->depth = 1;
|
|
|
|
/* uncompressed size */
|
|
|
|
entry->size = entry->delta_size = size;
|
|
|
|
entry->delta = base_entry;
|
|
|
|
entry->type = OBJ_DELTA;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* Otherwise we would do the usual */
|
2005-06-27 18:34:06 +08:00
|
|
|
}
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
|
|
|
|
if (sha1_object_info(entry->sha1, type, &entry->size))
|
2005-06-27 18:34:06 +08:00
|
|
|
die("unable to get type of object %s",
|
|
|
|
sha1_to_hex(entry->sha1));
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
|
|
|
|
if (!strcmp(type, "commit")) {
|
|
|
|
entry->type = OBJ_COMMIT;
|
|
|
|
} else if (!strcmp(type, "tree")) {
|
|
|
|
entry->type = OBJ_TREE;
|
|
|
|
} else if (!strcmp(type, "blob")) {
|
|
|
|
entry->type = OBJ_BLOB;
|
|
|
|
} else if (!strcmp(type, "tag")) {
|
|
|
|
entry->type = OBJ_TAG;
|
|
|
|
} else
|
|
|
|
die("unable to pack object %s of type %s",
|
|
|
|
sha1_to_hex(entry->sha1), type);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void hash_objects(void)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
struct object_entry *oe;
|
|
|
|
|
|
|
|
object_ix_hashsz = nr_objects * 2;
|
|
|
|
object_ix = xcalloc(sizeof(int), object_ix_hashsz);
|
|
|
|
for (i = 0, oe = objects; i < nr_objects; i++, oe++) {
|
|
|
|
int ix = locate_object_entry_hash(oe->sha1);
|
|
|
|
if (0 <= ix) {
|
|
|
|
error("the same object '%s' added twice",
|
|
|
|
sha1_to_hex(oe->sha1));
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
ix = -1 - ix;
|
|
|
|
object_ix[ix] = i + 1;
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void get_object_details(void)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
struct object_entry *entry = objects;
|
|
|
|
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
hash_objects();
|
|
|
|
prepare_pack_ix();
|
2005-06-26 05:42:43 +08:00
|
|
|
for (i = 0; i < nr_objects; i++)
|
|
|
|
check_object(entry++);
|
|
|
|
}
|
|
|
|
|
|
|
|
typedef int (*entry_sort_t)(const struct object_entry *, const struct object_entry *);
|
|
|
|
|
|
|
|
static entry_sort_t current_sort;
|
|
|
|
|
|
|
|
static int sort_comparator(const void *_a, const void *_b)
|
|
|
|
{
|
|
|
|
struct object_entry *a = *(struct object_entry **)_a;
|
|
|
|
struct object_entry *b = *(struct object_entry **)_b;
|
|
|
|
return current_sort(a,b);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct object_entry **create_sorted_list(entry_sort_t sort)
|
|
|
|
{
|
|
|
|
struct object_entry **list = xmalloc(nr_objects * sizeof(struct object_entry *));
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < nr_objects; i++)
|
|
|
|
list[i] = objects + i;
|
|
|
|
current_sort = sort;
|
|
|
|
qsort(list, nr_objects, sizeof(struct object_entry *), sort_comparator);
|
|
|
|
return list;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int sha1_sort(const struct object_entry *a, const struct object_entry *b)
|
|
|
|
{
|
|
|
|
return memcmp(a->sha1, b->sha1, 20);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int type_size_sort(const struct object_entry *a, const struct object_entry *b)
|
|
|
|
{
|
|
|
|
if (a->type < b->type)
|
|
|
|
return -1;
|
|
|
|
if (a->type > b->type)
|
|
|
|
return 1;
|
2005-06-27 06:27:28 +08:00
|
|
|
if (a->hash < b->hash)
|
|
|
|
return -1;
|
|
|
|
if (a->hash > b->hash)
|
|
|
|
return 1;
|
2005-06-26 05:42:43 +08:00
|
|
|
if (a->size < b->size)
|
|
|
|
return -1;
|
|
|
|
if (a->size > b->size)
|
|
|
|
return 1;
|
|
|
|
return a < b ? -1 : (a > b);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct unpacked {
|
|
|
|
struct object_entry *entry;
|
|
|
|
void *data;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
2005-06-27 04:43:41 +08:00
|
|
|
* We search for deltas _backwards_ in a list sorted by type and
|
|
|
|
* by size, so that we see progressively smaller and smaller files.
|
|
|
|
* That's because we prefer deltas to be from the bigger file
|
|
|
|
* to the smaller - deletes are potentially cheaper, but perhaps
|
|
|
|
* more importantly, the bigger file is likely the more recent
|
|
|
|
* one.
|
2005-06-26 05:42:43 +08:00
|
|
|
*/
|
2005-06-26 11:17:59 +08:00
|
|
|
static int try_delta(struct unpacked *cur, struct unpacked *old, unsigned max_depth)
|
2005-06-26 05:42:43 +08:00
|
|
|
{
|
|
|
|
struct object_entry *cur_entry = cur->entry;
|
|
|
|
struct object_entry *old_entry = old->entry;
|
2005-06-27 04:43:41 +08:00
|
|
|
unsigned long size, oldsize, delta_size, sizediff;
|
2005-06-26 10:30:20 +08:00
|
|
|
long max_size;
|
2005-06-26 05:42:43 +08:00
|
|
|
void *delta_buf;
|
|
|
|
|
|
|
|
/* Don't bother doing diffs between different types */
|
|
|
|
if (cur_entry->type != old_entry->type)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
size = cur_entry->size;
|
2005-06-26 10:30:20 +08:00
|
|
|
if (size < 50)
|
|
|
|
return -1;
|
2005-06-26 05:42:43 +08:00
|
|
|
oldsize = old_entry->size;
|
2005-06-27 04:43:41 +08:00
|
|
|
sizediff = oldsize > size ? oldsize - size : size - oldsize;
|
|
|
|
if (sizediff > size / 8)
|
2005-06-26 05:42:43 +08:00
|
|
|
return -1;
|
2005-06-26 11:17:59 +08:00
|
|
|
if (old_entry->depth >= max_depth)
|
|
|
|
return 0;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* NOTE!
|
|
|
|
*
|
|
|
|
* We always delta from the bigger to the smaller, since that's
|
|
|
|
* more space-efficient (deletes don't have to say _what_ they
|
|
|
|
* delete).
|
|
|
|
*/
|
2005-06-26 10:30:20 +08:00
|
|
|
max_size = size / 2 - 20;
|
|
|
|
if (cur_entry->delta)
|
|
|
|
max_size = cur_entry->delta_size-1;
|
2005-06-27 06:27:28 +08:00
|
|
|
if (sizediff >= max_size)
|
|
|
|
return -1;
|
2005-06-26 19:29:18 +08:00
|
|
|
delta_buf = diff_delta(old->data, oldsize,
|
|
|
|
cur->data, size, &delta_size, max_size);
|
2005-06-26 05:42:43 +08:00
|
|
|
if (!delta_buf)
|
2005-06-26 10:30:20 +08:00
|
|
|
return 0;
|
|
|
|
cur_entry->delta = old_entry;
|
|
|
|
cur_entry->delta_size = delta_size;
|
2005-06-26 11:17:59 +08:00
|
|
|
cur_entry->depth = old_entry->depth + 1;
|
2005-06-26 05:42:43 +08:00
|
|
|
free(delta_buf);
|
2005-06-26 08:36:26 +08:00
|
|
|
return 0;
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
2005-06-26 11:17:59 +08:00
|
|
|
static void find_deltas(struct object_entry **list, int window, int depth)
|
2005-06-26 05:42:43 +08:00
|
|
|
{
|
2005-06-27 04:43:41 +08:00
|
|
|
int i, idx;
|
2005-06-26 05:42:43 +08:00
|
|
|
unsigned int array_size = window * sizeof(struct unpacked);
|
|
|
|
struct unpacked *array = xmalloc(array_size);
|
2006-02-12 09:54:18 +08:00
|
|
|
int eye_candy;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
|
|
|
memset(array, 0, array_size);
|
2005-06-27 04:43:41 +08:00
|
|
|
i = nr_objects;
|
|
|
|
idx = 0;
|
2006-02-12 09:54:18 +08:00
|
|
|
eye_candy = i - (nr_objects / 20);
|
|
|
|
|
2005-06-27 04:43:41 +08:00
|
|
|
while (--i >= 0) {
|
2005-06-26 05:42:43 +08:00
|
|
|
struct object_entry *entry = list[i];
|
|
|
|
struct unpacked *n = array + idx;
|
|
|
|
unsigned long size;
|
|
|
|
char type[10];
|
|
|
|
int j;
|
|
|
|
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress && i <= eye_candy) {
|
|
|
|
eye_candy -= nr_objects / 20;
|
|
|
|
fputc('.', stderr);
|
|
|
|
}
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
|
|
|
|
if (entry->delta)
|
|
|
|
/* This happens if we decided to reuse existing
|
|
|
|
* delta from a pack.
|
|
|
|
*/
|
|
|
|
continue;
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
free(n->data);
|
|
|
|
n->entry = entry;
|
|
|
|
n->data = read_sha1_file(entry->sha1, type, &size);
|
|
|
|
if (size != entry->size)
|
|
|
|
die("object %s inconsistent object length (%lu vs %lu)", sha1_to_hex(entry->sha1), size, entry->size);
|
2005-06-26 09:29:23 +08:00
|
|
|
j = window;
|
|
|
|
while (--j > 0) {
|
|
|
|
unsigned int other_idx = idx + j;
|
2005-06-26 05:42:43 +08:00
|
|
|
struct unpacked *m;
|
2005-06-26 09:29:23 +08:00
|
|
|
if (other_idx >= window)
|
|
|
|
other_idx -= window;
|
2005-06-26 05:42:43 +08:00
|
|
|
m = array + other_idx;
|
|
|
|
if (!m->entry)
|
|
|
|
break;
|
2005-06-26 11:17:59 +08:00
|
|
|
if (try_delta(n, m, depth) < 0)
|
2005-06-26 05:42:43 +08:00
|
|
|
break;
|
|
|
|
}
|
2005-06-27 04:43:41 +08:00
|
|
|
idx++;
|
|
|
|
if (idx >= window)
|
|
|
|
idx = 0;
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
2005-08-09 02:46:58 +08:00
|
|
|
|
|
|
|
for (i = 0; i < window; ++i)
|
|
|
|
free(array[i].data);
|
|
|
|
free(array);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
|
|
|
|
2005-10-22 16:28:13 +08:00
|
|
|
static void prepare_pack(int window, int depth)
|
|
|
|
{
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress)
|
|
|
|
fprintf(stderr, "Packing %d objects", nr_objects);
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
get_object_details();
|
|
|
|
if (progress)
|
|
|
|
fprintf(stderr, ".");
|
|
|
|
|
2005-10-22 16:28:13 +08:00
|
|
|
sorted_by_type = create_sorted_list(type_size_sort);
|
|
|
|
if (window && depth)
|
|
|
|
find_deltas(sorted_by_type, window+1, depth);
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress)
|
|
|
|
fputc('\n', stderr);
|
2005-10-22 16:28:13 +08:00
|
|
|
write_pack_file();
|
|
|
|
}
|
|
|
|
|
|
|
|
static int reuse_cached_pack(unsigned char *sha1, int pack_to_stdout)
|
|
|
|
{
|
|
|
|
static const char cache[] = "pack-cache/pack-%s.%s";
|
|
|
|
char *cached_pack, *cached_idx;
|
|
|
|
int ifd, ofd, ifd_ix = -1;
|
|
|
|
|
|
|
|
cached_pack = git_path(cache, sha1_to_hex(sha1), "pack");
|
|
|
|
ifd = open(cached_pack, O_RDONLY);
|
|
|
|
if (ifd < 0)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (!pack_to_stdout) {
|
|
|
|
cached_idx = git_path(cache, sha1_to_hex(sha1), "idx");
|
|
|
|
ifd_ix = open(cached_idx, O_RDONLY);
|
|
|
|
if (ifd_ix < 0) {
|
|
|
|
close(ifd);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(stderr, "Reusing %d objects pack %s\n", nr_objects,
|
|
|
|
sha1_to_hex(sha1));
|
|
|
|
|
|
|
|
if (pack_to_stdout) {
|
|
|
|
if (copy_fd(ifd, 1))
|
|
|
|
exit(1);
|
|
|
|
close(ifd);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
char name[PATH_MAX];
|
|
|
|
snprintf(name, sizeof(name),
|
|
|
|
"%s-%s.%s", base_name, sha1_to_hex(sha1), "pack");
|
|
|
|
ofd = open(name, O_CREAT | O_EXCL | O_WRONLY, 0666);
|
|
|
|
if (ofd < 0)
|
|
|
|
die("unable to open %s (%s)", name, strerror(errno));
|
|
|
|
if (copy_fd(ifd, ofd))
|
|
|
|
exit(1);
|
|
|
|
close(ifd);
|
|
|
|
|
|
|
|
snprintf(name, sizeof(name),
|
|
|
|
"%s-%s.%s", base_name, sha1_to_hex(sha1), "idx");
|
|
|
|
ofd = open(name, O_CREAT | O_EXCL | O_WRONLY, 0666);
|
|
|
|
if (ofd < 0)
|
|
|
|
die("unable to open %s (%s)", name, strerror(errno));
|
|
|
|
if (copy_fd(ifd_ix, ofd))
|
|
|
|
exit(1);
|
|
|
|
close(ifd_ix);
|
|
|
|
puts(sha1_to_hex(sha1));
|
|
|
|
}
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
int main(int argc, char **argv)
|
|
|
|
{
|
2005-07-04 06:34:04 +08:00
|
|
|
SHA_CTX ctx;
|
2005-06-27 06:27:28 +08:00
|
|
|
char line[PATH_MAX + 20];
|
2005-06-29 02:10:48 +08:00
|
|
|
int window = 10, depth = 10, pack_to_stdout = 0;
|
2005-10-13 07:54:19 +08:00
|
|
|
struct object_entry **list;
|
2005-06-26 05:42:43 +08:00
|
|
|
int i;
|
2006-02-12 09:54:18 +08:00
|
|
|
struct timeval prev_tv;
|
|
|
|
int eye_candy = 0;
|
|
|
|
int eye_candy_incr = 500;
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
|
2005-11-26 16:50:02 +08:00
|
|
|
setup_git_directory();
|
|
|
|
|
2005-06-26 05:42:43 +08:00
|
|
|
for (i = 1; i < argc; i++) {
|
|
|
|
const char *arg = argv[i];
|
|
|
|
|
|
|
|
if (*arg == '-') {
|
2005-07-04 04:36:58 +08:00
|
|
|
if (!strcmp("--non-empty", arg)) {
|
|
|
|
non_empty = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2005-10-14 06:38:28 +08:00
|
|
|
if (!strcmp("--local", arg)) {
|
|
|
|
local = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2005-07-04 04:08:40 +08:00
|
|
|
if (!strcmp("--incremental", arg)) {
|
|
|
|
incremental = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
if (!strncmp("--window=", arg, 9)) {
|
|
|
|
char *end;
|
2005-06-26 10:35:47 +08:00
|
|
|
window = strtoul(arg+9, &end, 0);
|
2005-06-26 05:42:43 +08:00
|
|
|
if (!arg[9] || *end)
|
|
|
|
usage(pack_usage);
|
|
|
|
continue;
|
|
|
|
}
|
2005-06-26 11:17:59 +08:00
|
|
|
if (!strncmp("--depth=", arg, 8)) {
|
|
|
|
char *end;
|
|
|
|
depth = strtoul(arg+8, &end, 0);
|
|
|
|
if (!arg[8] || *end)
|
|
|
|
usage(pack_usage);
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-13 05:01:54 +08:00
|
|
|
if (!strcmp("-q", arg)) {
|
|
|
|
progress = 0;
|
|
|
|
continue;
|
|
|
|
}
|
2005-06-29 02:10:48 +08:00
|
|
|
if (!strcmp("--stdout", arg)) {
|
|
|
|
pack_to_stdout = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
usage(pack_usage);
|
|
|
|
}
|
|
|
|
if (base_name)
|
|
|
|
usage(pack_usage);
|
|
|
|
base_name = arg;
|
|
|
|
}
|
|
|
|
|
2005-06-29 02:10:48 +08:00
|
|
|
if (pack_to_stdout != !base_name)
|
2005-06-26 05:42:43 +08:00
|
|
|
usage(pack_usage);
|
|
|
|
|
2005-10-14 06:38:28 +08:00
|
|
|
prepare_packed_git();
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress) {
|
|
|
|
fprintf(stderr, "Generating pack...\n");
|
|
|
|
gettimeofday(&prev_tv, NULL);
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
while (fgets(line, sizeof(line), stdin) != NULL) {
|
2005-06-27 06:27:28 +08:00
|
|
|
unsigned int hash;
|
|
|
|
char *p;
|
2005-06-26 05:42:43 +08:00
|
|
|
unsigned char sha1[20];
|
2005-06-27 06:27:28 +08:00
|
|
|
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress && (eye_candy <= nr_objects)) {
|
|
|
|
fprintf(stderr, "Counting objects...%d\r", nr_objects);
|
|
|
|
if (eye_candy && (50 <= eye_candy_incr)) {
|
|
|
|
struct timeval tv;
|
|
|
|
int time_diff;
|
|
|
|
gettimeofday(&tv, NULL);
|
|
|
|
time_diff = (tv.tv_sec - prev_tv.tv_sec);
|
|
|
|
time_diff <<= 10;
|
|
|
|
time_diff += (tv.tv_usec - prev_tv.tv_usec);
|
|
|
|
if ((1 << 9) < time_diff)
|
|
|
|
eye_candy_incr += 50;
|
|
|
|
else if (50 < eye_candy_incr)
|
|
|
|
eye_candy_incr -= 50;
|
|
|
|
}
|
|
|
|
eye_candy += eye_candy_incr;
|
|
|
|
}
|
2005-06-26 05:42:43 +08:00
|
|
|
if (get_sha1_hex(line, sha1))
|
2005-11-22 04:38:31 +08:00
|
|
|
die("expected sha1, got garbage:\n %s", line);
|
2005-06-27 06:27:28 +08:00
|
|
|
hash = 0;
|
|
|
|
p = line+40;
|
|
|
|
while (*p) {
|
|
|
|
unsigned char c = *p++;
|
|
|
|
if (isspace(c))
|
|
|
|
continue;
|
|
|
|
hash = hash * 11 + c;
|
|
|
|
}
|
2005-10-13 07:54:19 +08:00
|
|
|
add_object_entry(sha1, hash);
|
2005-06-26 05:42:43 +08:00
|
|
|
}
|
2006-02-12 09:54:18 +08:00
|
|
|
if (progress)
|
|
|
|
fprintf(stderr, "Done counting %d objects.\n", nr_objects);
|
2005-07-04 04:36:58 +08:00
|
|
|
if (non_empty && !nr_objects)
|
|
|
|
return 0;
|
2005-06-26 05:42:43 +08:00
|
|
|
|
|
|
|
sorted_by_sha = create_sorted_list(sha1_sort);
|
2005-10-13 07:54:19 +08:00
|
|
|
SHA1_Init(&ctx);
|
|
|
|
list = sorted_by_sha;
|
|
|
|
for (i = 0; i < nr_objects; i++) {
|
|
|
|
struct object_entry *entry = *list++;
|
|
|
|
SHA1_Update(&ctx, entry->sha1, 20);
|
|
|
|
}
|
|
|
|
SHA1_Final(object_list_sha1, &ctx);
|
|
|
|
|
2005-10-22 16:28:13 +08:00
|
|
|
if (reuse_cached_pack(object_list_sha1, pack_to_stdout))
|
|
|
|
;
|
|
|
|
else {
|
|
|
|
prepare_pack(window, depth);
|
|
|
|
if (!pack_to_stdout) {
|
|
|
|
write_index_file();
|
|
|
|
puts(sha1_to_hex(object_list_sha1));
|
|
|
|
}
|
2005-07-04 06:34:04 +08:00
|
|
|
}
|
pack-objects: reuse data from existing packs.
When generating a new pack, notice if we have already needed
objects in existing packs. If an object is stored deltified,
and its base object is also what we are going to pack, then
reuse the existing deltified representation unconditionally,
bypassing all the expensive find_deltas() and try_deltas()
calls.
Also, notice if what we are going to write out exactly match
what is already in an existing pack (either deltified or just
compressed). In such a case, we can just copy it instead of
going through the usual uncompressing & recompressing cycle.
Without this patch, in linux-2.6 repository with about 1500
loose objects and a single mega pack:
$ git-rev-list --objects v2.6.16-rc3 >RL
$ wc -l RL
184141 RL
$ time git-pack-objects p <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
real 12m4.323s
user 11m2.560s
sys 0m55.950s
With this patch, the same input:
$ time ../git.junio/git-pack-objects q <RL
Generating pack...
Done counting 184141 objects.
Packing 184141 objects.....................
a1fc7b3e537fcb9b3c46b7505df859f0a11e79d2
Total 184141, written 184141, reused 182441
real 1m2.608s
user 0m55.090s
sys 0m1.830s
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-02-16 09:34:29 +08:00
|
|
|
fprintf(stderr, "Total %d, written %d, reused %d\n",
|
|
|
|
nr_objects, written, reused);
|
2005-06-26 05:42:43 +08:00
|
|
|
return 0;
|
|
|
|
}
|