#include "delta.h"
#include "pack.h"
#include "csum-file.h"
+#include "diff.h"
#include <sys/time.h>
#include <signal.h>
struct object_entry *delta_sibling; /* other deltified objects who
* uses the same base as me
*/
+ int preferred_base; /* we do not pack this, but is encouraged to
+ * be used as the base objectto delta huge
+ * objects against.
+ */
};
/*
static int incremental = 0;
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 int nr_objects = 0, nr_alloc = 0, nr_result = 0;
static const char *base_name;
static unsigned char pack_file_sha1[20];
static int progress = 1;
static int pack_revindex_ix(struct packed_git *p)
{
- unsigned int ui = (unsigned int) p;
+ unsigned long ui = (unsigned long)p;
int i;
ui = ui ^ (ui >> 16); /* defeat structure alignment */
return n;
}
-static unsigned long write_object(struct sha1file *f, struct object_entry *entry)
+static unsigned long write_object(struct sha1file *f,
+ struct object_entry *entry)
{
unsigned long size;
char type[10];
enum object_type obj_type;
int to_reuse = 0;
+ if (entry->preferred_base)
+ return 0;
+
obj_type = entry->type;
if (! entry->in_pack)
to_reuse = 0; /* can't reuse what we don't have */
struct sha1file *f;
unsigned long offset;
struct pack_header hdr;
+ unsigned last_percent = 999;
+ int do_progress = 0;
if (!base_name)
f = sha1fd(1, "<stdout>");
- else
- f = sha1create("%s-%s.%s", base_name, sha1_to_hex(object_list_sha1), "pack");
+ else {
+ f = sha1create("%s-%s.%s", base_name,
+ sha1_to_hex(object_list_sha1), "pack");
+ do_progress = progress;
+ }
+ if (do_progress)
+ fprintf(stderr, "Writing %d objects.\n", nr_result);
+
hdr.hdr_signature = htonl(PACK_SIGNATURE);
hdr.hdr_version = htonl(PACK_VERSION);
- hdr.hdr_entries = htonl(nr_objects);
+ hdr.hdr_entries = htonl(nr_result);
sha1write(f, &hdr, sizeof(hdr));
offset = sizeof(hdr);
+ if (!nr_result)
+ goto done;
for (i = 0; i < nr_objects; i++) {
offset = write_one(f, objects + i, offset);
- if (progress_update) {
- fprintf(stderr, "Writing (%d %d%%)\r",
- i+1, (i+1) * 100/nr_objects);
- progress_update = 0;
+ if (do_progress) {
+ unsigned percent = written * 100 / nr_result;
+ if (progress_update || percent != last_percent) {
+ fprintf(stderr, "%4u%% (%u/%u) done\r",
+ percent, written, nr_result);
+ progress_update = 0;
+ last_percent = percent;
+ }
}
}
-
+ if (do_progress)
+ fputc('\n', stderr);
+ done:
sha1close(f, pack_file_sha1, 1);
}
static void write_index_file(void)
{
int i;
- struct sha1file *f = sha1create("%s-%s.%s", base_name, sha1_to_hex(object_list_sha1), "idx");
+ struct sha1file *f = sha1create("%s-%s.%s", base_name,
+ sha1_to_hex(object_list_sha1), "idx");
struct object_entry **list = sorted_by_sha;
- struct object_entry **last = list + nr_objects;
+ struct object_entry **last = list + nr_result;
unsigned int array[256];
/*
* Write the actual SHA1 entries..
*/
list = sorted_by_sha;
- for (i = 0; i < nr_objects; i++) {
+ for (i = 0; i < nr_result; i++) {
struct object_entry *entry = *list++;
unsigned int offset = htonl(entry->offset);
sha1write(f, &offset, 4);
sha1close(f, NULL, 1);
}
-static int add_object_entry(unsigned char *sha1, unsigned int hash)
+static int locate_object_entry_hash(const 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(const unsigned char *sha1)
+{
+ int i;
+
+ if (!object_ix_hashsz)
+ return NULL;
+
+ i = locate_object_entry_hash(sha1);
+ if (0 <= i)
+ return &objects[object_ix[i]-1];
+ return NULL;
+}
+
+static void rehash_objects(void)
+{
+ int i;
+ struct object_entry *oe;
+
+ object_ix_hashsz = nr_objects * 3;
+ if (object_ix_hashsz < 1024)
+ object_ix_hashsz = 1024;
+ object_ix = xrealloc(object_ix, sizeof(int) * object_ix_hashsz);
+ object_ix = memset(object_ix, 0, 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)
+ continue;
+ ix = -1 - ix;
+ object_ix[ix] = i + 1;
+ }
+}
+
+struct name_path {
+ struct name_path *up;
+ const char *elem;
+ int len;
+};
+
+#define DIRBITS 12
+
+static unsigned name_hash(struct name_path *path, const char *name)
+{
+ struct name_path *p = path;
+ const char *n = name + strlen(name);
+ unsigned hash = 0, name_hash = 0, name_done = 0;
+
+ if (n != name && n[-1] == '\n')
+ n--;
+ while (name <= --n) {
+ unsigned char c = *n;
+ if (c == '/' && !name_done) {
+ name_hash = hash;
+ name_done = 1;
+ hash = 0;
+ }
+ hash = hash * 11 + c;
+ }
+ if (!name_done) {
+ name_hash = hash;
+ hash = 0;
+ }
+ for (p = path; p; p = p->up) {
+ hash = hash * 11 + '/';
+ n = p->elem + p->len;
+ while (p->elem <= --n) {
+ unsigned char c = *n;
+ hash = hash * 11 + c;
+ }
+ }
+ /*
+ * Make sure "Makefile" and "t/Makefile" are hashed separately
+ * but close enough.
+ */
+ hash = (name_hash<<DIRBITS) | (hash & ((1U<<DIRBITS )-1));
+
+ if (0) { /* debug */
+ n = name + strlen(name);
+ if (n != name && n[-1] == '\n')
+ n--;
+ while (name <= --n)
+ fputc(*n, stderr);
+ for (p = path; p; p = p->up) {
+ fputc('/', stderr);
+ n = p->elem + p->len;
+ while (p->elem <= --n)
+ fputc(*n, stderr);
+ }
+ fprintf(stderr, "\t%08x\n", hash);
+ }
+ return hash;
+}
+
+static int add_object_entry(const unsigned char *sha1, unsigned hash, int exclude)
{
unsigned int idx = nr_objects;
struct object_entry *entry;
struct packed_git *p;
unsigned int found_offset = 0;
struct packed_git *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;
+ int ix, status = 0;
+
+ if (!exclude) {
+ 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;
+ }
}
}
}
+ if ((entry = locate_object_entry(sha1)) != NULL)
+ goto already_added;
if (idx >= nr_alloc) {
unsigned int needed = (idx + 1024) * 3 / 2;
nr_alloc = needed;
}
entry = objects + idx;
+ nr_objects = idx + 1;
memset(entry, 0, sizeof(*entry));
memcpy(entry->sha1, sha1, 20);
entry->hash = hash;
- if (found_pack) {
- entry->in_pack = found_pack;
- entry->in_pack_offset = found_offset;
+
+ if (object_ix_hashsz * 3 <= nr_objects * 4)
+ rehash_objects();
+ else {
+ ix = locate_object_entry_hash(entry->sha1);
+ if (0 <= ix)
+ die("internal error in object hashing.");
+ object_ix[-1 - ix] = idx + 1;
}
- nr_objects = idx+1;
- return 1;
+ status = 1;
+
+ already_added:
+ if (progress_update) {
+ fprintf(stderr, "Counting objects...%d\r", nr_objects);
+ progress_update = 0;
+ }
+ if (exclude)
+ entry->preferred_base = 1;
+ else {
+ if (found_pack) {
+ entry->in_pack = found_pack;
+ entry->in_pack_offset = found_offset;
+ }
+ }
+ return status;
}
-static int locate_object_entry_hash(unsigned char *sha1)
+static void add_pbase_tree(struct tree_desc *tree, struct name_path *up)
{
- 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;
+ while (tree->size) {
+ const unsigned char *sha1;
+ const char *name;
+ unsigned mode, hash;
+ unsigned long size;
+ char type[20];
+
+ sha1 = tree_entry_extract(tree, &name, &mode);
+ update_tree_entry(tree);
+ if (!has_sha1_file(sha1))
+ continue;
+ if (sha1_object_info(sha1, type, &size))
+ continue;
+
+ hash = name_hash(up, name);
+ if (!add_object_entry(sha1, hash, 1))
+ continue;
+
+ if (!strcmp(type, "tree")) {
+ struct tree_desc sub;
+ void *elem;
+ struct name_path me;
+
+ elem = read_sha1_file(sha1, type, &sub.size);
+ sub.buf = elem;
+ if (sub.buf) {
+ me.up = up;
+ me.elem = name;
+ me.len = strlen(name);
+ add_pbase_tree(&sub, &me);
+ free(elem);
+ }
+ }
}
- return -1 - i;
}
-static struct object_entry *locate_object_entry(unsigned char *sha1)
+static void add_preferred_base(unsigned char *sha1)
{
- int i = locate_object_entry_hash(sha1);
- if (0 <= i)
- return &objects[object_ix[i]-1];
- return NULL;
+ struct tree_desc tree;
+ void *elem;
+
+ elem = read_object_with_reference(sha1, "tree", &tree.size, NULL);
+ tree.buf = elem;
+ if (!tree.buf)
+ return;
+ if (add_object_entry(sha1, name_hash(NULL, ""), 1))
+ add_pbase_tree(&tree, NULL);
+ free(elem);
}
static void check_object(struct object_entry *entry)
{
char type[20];
- if (entry->in_pack) {
+ if (entry->in_pack && !entry->preferred_base) {
unsigned char base[20];
unsigned long size;
struct object_entry *base_entry;
*/
if (!no_reuse_delta &&
entry->in_pack_type == OBJ_DELTA &&
- (base_entry = locate_object_entry(base))) {
+ (base_entry = locate_object_entry(base)) &&
+ (!base_entry->preferred_base)) {
/* Depth value does not matter - find_deltas()
* will never consider reused delta as the
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;
- }
-}
-
static unsigned int check_delta_limit(struct object_entry *me, unsigned int n)
{
struct object_entry *child = me->delta_child;
int i;
struct object_entry *entry;
- hash_objects();
prepare_pack_ix();
for (i = 0, entry = objects; i < nr_objects; i++, entry++)
check_object(entry);
- for (i = 0, entry = objects; i < nr_objects; i++, entry++)
- if (!entry->delta && entry->delta_child)
- entry->delta_limit =
- check_delta_limit(entry, 1);
+
+ if (nr_objects == nr_result) {
+ /*
+ * Depth of objects that depend on the entry -- this
+ * is subtracted from depth-max to break too deep
+ * delta chain because of delta data reusing.
+ * However, we loosen this restriction when we know we
+ * are creating a thin pack -- it will have to be
+ * expanded on the other end anyway, so do not
+ * artificially cut the delta chain and let it go as
+ * deep as it wants.
+ */
+ for (i = 0, entry = objects; i < nr_objects; i++, entry++)
+ if (!entry->delta && entry->delta_child)
+ entry->delta_limit =
+ check_delta_limit(entry, 1);
+ }
}
typedef int (*entry_sort_t)(const struct object_entry *, const struct object_entry *);
return memcmp(a->sha1, b->sha1, 20);
}
+static struct object_entry **create_final_object_list(void)
+{
+ struct object_entry **list;
+ int i, j;
+
+ for (i = nr_result = 0; i < nr_objects; i++)
+ if (!objects[i].preferred_base)
+ nr_result++;
+ list = xmalloc(nr_result * sizeof(struct object_entry *));
+ for (i = j = 0; i < nr_objects; i++) {
+ if (!objects[i].preferred_base)
+ list[j++] = objects + i;
+ }
+ current_sort = sha1_sort;
+ qsort(list, nr_result, sizeof(struct object_entry *), sort_comparator);
+ return list;
+}
+
static int type_size_sort(const struct object_entry *a, const struct object_entry *b)
{
if (a->type < b->type)
return -1;
if (a->hash > b->hash)
return 1;
+ if (a->preferred_base < b->preferred_base)
+ return -1;
+ if (a->preferred_base > b->preferred_base)
+ return 1;
if (a->size < b->size)
return -1;
if (a->size > b->size)
if (cur_entry->type != old_entry->type)
return -1;
- /* If the current object is at edge, take the depth the objects
- * that depend on the current object into account -- otherwise
- * they would become too deep.
+ /* We do not compute delta to *create* objects we are not
+ * going to pack.
+ */
+ if (cur_entry->preferred_base)
+ return -1;
+
+ /* If the current object is at pack edge, take the depth the
+ * objects that depend on the current object into account --
+ * otherwise they would become too deep.
*/
if (cur_entry->delta_child) {
if (max_depth <= cur_entry->delta_limit)
}
size = cur_entry->size;
- if (size < 50)
- return -1;
oldsize = old_entry->size;
sizediff = oldsize > size ? oldsize - size : size - oldsize;
- if (sizediff > size / 8)
+
+ if (size < 50)
return -1;
if (old_entry->depth >= max_depth)
return 0;
int i, idx;
unsigned int array_size = window * sizeof(struct unpacked);
struct unpacked *array = xmalloc(array_size);
+ unsigned processed = 0;
+ unsigned last_percent = 999;
memset(array, 0, array_size);
i = nr_objects;
idx = 0;
+ if (progress)
+ fprintf(stderr, "Deltifying %d objects.\n", nr_result);
while (--i >= 0) {
struct object_entry *entry = list[i];
char type[10];
int j;
- if (progress_update || i == 0) {
- fprintf(stderr, "Deltifying (%d %d%%)\r",
- nr_objects-i, (nr_objects-i) * 100/nr_objects);
- progress_update = 0;
+ if (!entry->preferred_base)
+ processed++;
+
+ if (progress) {
+ unsigned percent = processed * 100 / nr_result;
+ if (percent != last_percent || progress_update) {
+ fprintf(stderr, "%4u%% (%u/%u) done\r",
+ percent, processed, nr_result);
+ progress_update = 0;
+ last_percent = percent;
+ }
}
if (entry->delta)
if (try_delta(n, m, depth) < 0)
break;
}
+#if 0
+ /* if we made n a delta, and if n is already at max
+ * depth, leaving it in the window is pointless. we
+ * should evict it first.
+ * ... in theory only; somehow this makes things worse.
+ */
+ if (entry->delta && depth <= entry->depth)
+ continue;
+#endif
idx++;
if (idx >= window)
idx = 0;
}
while (fgets(line, sizeof(line), stdin) != NULL) {
- unsigned int hash;
- char *p;
unsigned char sha1[20];
- if (progress_update) {
- fprintf(stderr, "Counting objects...%d\r", nr_objects);
- progress_update = 0;
+ if (line[0] == '-') {
+ if (get_sha1_hex(line+1, sha1))
+ die("expected edge sha1, got garbage:\n %s",
+ line+1);
+ add_preferred_base(sha1);
+ continue;
}
if (get_sha1_hex(line, sha1))
die("expected sha1, got garbage:\n %s", line);
- hash = 0;
- p = line+40;
- while (*p) {
- unsigned char c = *p++;
- if (isspace(c))
- continue;
- hash = hash * 11 + c;
- }
- add_object_entry(sha1, hash);
+ add_object_entry(sha1, name_hash(NULL, line+41), 0);
}
if (progress)
fprintf(stderr, "Done counting %d objects.\n", nr_objects);
- if (non_empty && !nr_objects)
+ sorted_by_sha = create_final_object_list();
+ if (non_empty && !nr_result)
return 0;
- sorted_by_sha = create_sorted_list(sha1_sort);
SHA1_Init(&ctx);
list = sorted_by_sha;
- for (i = 0; i < nr_objects; i++) {
+ for (i = 0; i < nr_result; i++) {
struct object_entry *entry = *list++;
SHA1_Update(&ctx, entry->sha1, 20);
}
SHA1_Final(object_list_sha1, &ctx);
+ if (progress && (nr_objects != nr_result))
+ fprintf(stderr, "Result has %d objects.\n", nr_result);
if (reuse_cached_pack(object_list_sha1, pack_to_stdout))
;
else {
- prepare_pack(window, depth);
+ if (nr_result)
+ prepare_pack(window, depth);
if (progress && pack_to_stdout) {
/* the other end usually displays progress itself */
struct itimerval v = {{0,},};
}
if (progress)
fprintf(stderr, "Total %d, written %d (delta %d), reused %d (delta %d)\n",
- nr_objects, written, written_delta, reused, reused_delta);
+ nr_result, written, written_delta, reused, reused_delta);
return 0;
}