2005-04-18 22:04:43 +02:00
|
|
|
/*
|
|
|
|
* GIT - The information manager from hell
|
|
|
|
*
|
|
|
|
* Copyright (C) Linus Torvalds, 2005
|
|
|
|
*
|
|
|
|
* This handles basic git sha1 object files - packing, unpacking,
|
|
|
|
* creation etc.
|
|
|
|
*/
|
|
|
|
#include "cache.h"
|
2005-06-27 12:35:33 +02:00
|
|
|
#include "delta.h"
|
2005-06-28 23:21:02 +02:00
|
|
|
#include "pack.h"
|
2006-04-02 14:44:09 +02:00
|
|
|
#include "blob.h"
|
|
|
|
#include "commit.h"
|
|
|
|
#include "tag.h"
|
|
|
|
#include "tree.h"
|
2005-04-18 22:04:43 +02:00
|
|
|
|
2005-04-23 20:09:32 +02:00
|
|
|
#ifndef O_NOATIME
|
|
|
|
#if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
|
|
|
|
#define O_NOATIME 01000000
|
|
|
|
#else
|
|
|
|
#define O_NOATIME 0
|
|
|
|
#endif
|
|
|
|
#endif
|
|
|
|
|
2006-08-15 19:23:48 +02:00
|
|
|
const unsigned char null_sha1[20];
|
2005-09-30 23:02:47 +02:00
|
|
|
|
2005-04-23 20:09:32 +02:00
|
|
|
static unsigned int sha1_file_open_flag = O_NOATIME;
|
|
|
|
|
2006-09-21 01:04:46 +02:00
|
|
|
signed char hexval_table[256] = {
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 00-07 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 08-0f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 10-17 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 18-1f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 20-27 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 28-2f */
|
|
|
|
0, 1, 2, 3, 4, 5, 6, 7, /* 30-37 */
|
|
|
|
8, 9, -1, -1, -1, -1, -1, -1, /* 38-3f */
|
|
|
|
-1, 10, 11, 12, 13, 14, 15, -1, /* 40-47 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 48-4f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 50-57 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 58-5f */
|
|
|
|
-1, 10, 11, 12, 13, 14, 15, -1, /* 60-67 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 68-67 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 70-77 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 78-7f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 80-87 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 88-8f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 90-97 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* 98-9f */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* a0-a7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* a8-af */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* b0-b7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* b8-bf */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* c0-c7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* c8-cf */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* d0-d7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* d8-df */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* e0-e7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* e8-ef */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* f0-f7 */
|
|
|
|
-1, -1, -1, -1, -1, -1, -1, -1, /* f8-ff */
|
|
|
|
};
|
2005-04-18 22:04:43 +02:00
|
|
|
|
|
|
|
int get_sha1_hex(const char *hex, unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < 20; i++) {
|
|
|
|
unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]);
|
|
|
|
if (val & ~0xff)
|
|
|
|
return -1;
|
|
|
|
*sha1++ = val;
|
|
|
|
hex += 2;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-07-06 10:11:52 +02:00
|
|
|
int safe_create_leading_directories(char *path)
|
|
|
|
{
|
|
|
|
char *pos = path;
|
2006-02-10 02:56:13 +01:00
|
|
|
struct stat st;
|
|
|
|
|
2005-11-07 00:36:15 +01:00
|
|
|
if (*pos == '/')
|
|
|
|
pos++;
|
2005-07-06 10:11:52 +02:00
|
|
|
|
|
|
|
while (pos) {
|
|
|
|
pos = strchr(pos, '/');
|
|
|
|
if (!pos)
|
|
|
|
break;
|
|
|
|
*pos = 0;
|
2006-02-10 02:56:13 +01:00
|
|
|
if (!stat(path, &st)) {
|
|
|
|
/* path exists */
|
|
|
|
if (!S_ISDIR(st.st_mode)) {
|
2005-07-06 10:11:52 +02:00
|
|
|
*pos = '/';
|
2006-02-10 02:56:13 +01:00
|
|
|
return -3;
|
2005-07-06 10:11:52 +02:00
|
|
|
}
|
2005-12-22 23:13:56 +01:00
|
|
|
}
|
2006-02-10 02:56:13 +01:00
|
|
|
else if (mkdir(path, 0777)) {
|
|
|
|
*pos = '/';
|
|
|
|
return -1;
|
|
|
|
}
|
2005-12-22 23:13:56 +01:00
|
|
|
else if (adjust_shared_perm(path)) {
|
|
|
|
*pos = '/';
|
|
|
|
return -2;
|
|
|
|
}
|
2005-07-06 10:11:52 +02:00
|
|
|
*pos++ = '/';
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2005-07-05 20:31:32 +02:00
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
char * sha1_to_hex(const unsigned char *sha1)
|
|
|
|
{
|
2006-05-04 02:21:08 +02:00
|
|
|
static int bufno;
|
|
|
|
static char hexbuffer[4][50];
|
2005-04-18 22:04:43 +02:00
|
|
|
static const char hex[] = "0123456789abcdef";
|
2006-05-04 02:21:08 +02:00
|
|
|
char *buffer = hexbuffer[3 & ++bufno], *buf = buffer;
|
2005-04-18 22:04:43 +02:00
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < 20; i++) {
|
|
|
|
unsigned int val = *sha1++;
|
|
|
|
*buf++ = hex[val >> 4];
|
|
|
|
*buf++ = hex[val & 0xf];
|
|
|
|
}
|
2005-12-22 18:55:59 +01:00
|
|
|
*buf = '\0';
|
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
return buffer;
|
|
|
|
}
|
|
|
|
|
2005-05-07 09:38:04 +02:00
|
|
|
static void fill_sha1_path(char *pathbuf, const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < 20; i++) {
|
|
|
|
static char hex[] = "0123456789abcdef";
|
|
|
|
unsigned int val = sha1[i];
|
|
|
|
char *pos = pathbuf + i*2 + (i > 0);
|
|
|
|
*pos++ = hex[val >> 4];
|
|
|
|
*pos = hex[val & 0xf];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
/*
|
|
|
|
* NOTE! This returns a statically allocated buffer, so you have to be
|
2006-09-02 06:16:31 +02:00
|
|
|
* careful about using it. Do a "xstrdup()" if you need to save the
|
2005-04-18 22:04:43 +02:00
|
|
|
* filename.
|
2005-05-07 09:38:04 +02:00
|
|
|
*
|
|
|
|
* Also note that this returns the location for creating. Reading
|
|
|
|
* SHA1 file can happen from any alternate directory listed in the
|
2005-05-10 02:57:56 +02:00
|
|
|
* DB_ENVIRONMENT environment variable if it is not found in
|
2005-05-07 09:38:04 +02:00
|
|
|
* the primary object database.
|
2005-04-18 22:04:43 +02:00
|
|
|
*/
|
|
|
|
char *sha1_file_name(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
static char *name, *base;
|
|
|
|
|
|
|
|
if (!base) {
|
2005-05-10 02:57:56 +02:00
|
|
|
const char *sha1_file_directory = get_object_directory();
|
2005-04-18 22:04:43 +02:00
|
|
|
int len = strlen(sha1_file_directory);
|
2005-04-26 21:00:58 +02:00
|
|
|
base = xmalloc(len + 60);
|
2005-04-18 22:04:43 +02:00
|
|
|
memcpy(base, sha1_file_directory, len);
|
|
|
|
memset(base+len, 0, 60);
|
|
|
|
base[len] = '/';
|
|
|
|
base[len+3] = '/';
|
|
|
|
name = base + len + 1;
|
|
|
|
}
|
2005-05-07 09:38:04 +02:00
|
|
|
fill_sha1_path(name, sha1);
|
2005-04-18 22:04:43 +02:00
|
|
|
return base;
|
|
|
|
}
|
|
|
|
|
2005-08-01 02:53:44 +02:00
|
|
|
char *sha1_pack_name(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
static const char hex[] = "0123456789abcdef";
|
|
|
|
static char *name, *base, *buf;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!base) {
|
|
|
|
const char *sha1_file_directory = get_object_directory();
|
|
|
|
int len = strlen(sha1_file_directory);
|
|
|
|
base = xmalloc(len + 60);
|
|
|
|
sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.pack", sha1_file_directory);
|
|
|
|
name = base + len + 11;
|
|
|
|
}
|
|
|
|
|
|
|
|
buf = name;
|
|
|
|
|
|
|
|
for (i = 0; i < 20; i++) {
|
|
|
|
unsigned int val = *sha1++;
|
|
|
|
*buf++ = hex[val >> 4];
|
|
|
|
*buf++ = hex[val & 0xf];
|
|
|
|
}
|
|
|
|
|
|
|
|
return base;
|
|
|
|
}
|
|
|
|
|
|
|
|
char *sha1_pack_index_name(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
static const char hex[] = "0123456789abcdef";
|
|
|
|
static char *name, *base, *buf;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!base) {
|
|
|
|
const char *sha1_file_directory = get_object_directory();
|
|
|
|
int len = strlen(sha1_file_directory);
|
|
|
|
base = xmalloc(len + 60);
|
|
|
|
sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.idx", sha1_file_directory);
|
|
|
|
name = base + len + 11;
|
|
|
|
}
|
|
|
|
|
|
|
|
buf = name;
|
|
|
|
|
|
|
|
for (i = 0; i < 20; i++) {
|
|
|
|
unsigned int val = *sha1++;
|
|
|
|
*buf++ = hex[val >> 4];
|
|
|
|
*buf++ = hex[val & 0xf];
|
|
|
|
}
|
|
|
|
|
|
|
|
return base;
|
|
|
|
}
|
|
|
|
|
2005-08-15 02:25:57 +02:00
|
|
|
struct alternate_object_database *alt_odb_list;
|
|
|
|
static struct alternate_object_database **alt_odb_tail;
|
2005-05-07 09:38:04 +02:00
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
static void read_info_alternates(const char * alternates, int depth);
|
|
|
|
|
2005-05-08 22:51:13 +02:00
|
|
|
/*
|
|
|
|
* Prepare alternate object database registry.
|
2005-08-15 02:25:57 +02:00
|
|
|
*
|
|
|
|
* The variable alt_odb_list points at the list of struct
|
|
|
|
* alternate_object_database. The elements on this list come from
|
|
|
|
* non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT
|
|
|
|
* environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates,
|
2005-12-05 07:48:43 +01:00
|
|
|
* whose contents is similar to that environment variable but can be
|
|
|
|
* LF separated. Its base points at a statically allocated buffer that
|
2005-08-15 02:25:57 +02:00
|
|
|
* contains "/the/directory/corresponding/to/.git/objects/...", while
|
|
|
|
* its name points just after the slash at the end of ".git/objects/"
|
|
|
|
* in the example above, and has enough space to hold 40-byte hex
|
|
|
|
* SHA1, an extra slash for the first level indirection, and the
|
|
|
|
* terminating NUL.
|
2005-05-08 22:51:13 +02:00
|
|
|
*/
|
2006-05-07 20:19:21 +02:00
|
|
|
static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth)
|
2005-05-07 09:38:04 +02:00
|
|
|
{
|
2006-05-07 20:19:21 +02:00
|
|
|
struct stat st;
|
2005-12-05 07:48:43 +01:00
|
|
|
const char *objdir = get_object_directory();
|
2006-05-07 20:19:21 +02:00
|
|
|
struct alternate_object_database *ent;
|
|
|
|
struct alternate_object_database *alt;
|
|
|
|
/* 43 = 40-byte + 2 '/' + terminating NUL */
|
|
|
|
int pfxlen = len;
|
|
|
|
int entlen = pfxlen + 43;
|
2005-09-13 09:05:22 +02:00
|
|
|
int base_len = -1;
|
2005-08-15 02:25:57 +02:00
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
if (*entry != '/' && relative_base) {
|
|
|
|
/* Relative alt-odb */
|
|
|
|
if (base_len < 0)
|
|
|
|
base_len = strlen(relative_base) + 1;
|
|
|
|
entlen += base_len;
|
|
|
|
pfxlen += base_len;
|
|
|
|
}
|
|
|
|
ent = xmalloc(sizeof(*ent) + entlen);
|
|
|
|
|
|
|
|
if (*entry != '/' && relative_base) {
|
|
|
|
memcpy(ent->base, relative_base, base_len - 1);
|
|
|
|
ent->base[base_len - 1] = '/';
|
|
|
|
memcpy(ent->base + base_len, entry, len);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
memcpy(ent->base, entry, pfxlen);
|
|
|
|
|
|
|
|
ent->name = ent->base + pfxlen + 1;
|
|
|
|
ent->base[pfxlen + 3] = '/';
|
|
|
|
ent->base[pfxlen] = ent->base[entlen-1] = 0;
|
|
|
|
|
|
|
|
/* Detect cases where alternate disappeared */
|
|
|
|
if (stat(ent->base, &st) || !S_ISDIR(st.st_mode)) {
|
|
|
|
error("object directory %s does not exist; "
|
|
|
|
"check .git/objects/info/alternates.",
|
|
|
|
ent->base);
|
|
|
|
free(ent);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Prevent the common mistake of listing the same
|
|
|
|
* thing twice, or object directory itself.
|
|
|
|
*/
|
|
|
|
for (alt = alt_odb_list; alt; alt = alt->next) {
|
|
|
|
if (!memcmp(ent->base, alt->base, pfxlen)) {
|
|
|
|
free(ent);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (!memcmp(ent->base, objdir, pfxlen)) {
|
|
|
|
free(ent);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* add the alternate entry */
|
|
|
|
*alt_odb_tail = ent;
|
|
|
|
alt_odb_tail = &(ent->next);
|
|
|
|
ent->next = NULL;
|
|
|
|
|
|
|
|
/* recursively add alternates */
|
|
|
|
read_info_alternates(ent->base, depth + 1);
|
|
|
|
|
|
|
|
ent->base[pfxlen] = '/';
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void link_alt_odb_entries(const char *alt, const char *ep, int sep,
|
|
|
|
const char *relative_base, int depth)
|
|
|
|
{
|
|
|
|
const char *cp, *last;
|
|
|
|
|
|
|
|
if (depth > 5) {
|
|
|
|
error("%s: ignoring alternate object stores, nesting too deep.",
|
|
|
|
relative_base);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2005-08-15 02:25:57 +02:00
|
|
|
last = alt;
|
2005-08-17 03:22:05 +02:00
|
|
|
while (last < ep) {
|
|
|
|
cp = last;
|
|
|
|
if (cp < ep && *cp == '#') {
|
|
|
|
while (cp < ep && *cp != sep)
|
|
|
|
cp++;
|
|
|
|
last = cp + 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-05-07 20:19:21 +02:00
|
|
|
while (cp < ep && *cp != sep)
|
|
|
|
cp++;
|
2005-08-15 02:25:57 +02:00
|
|
|
if (last != cp) {
|
2006-05-07 20:19:21 +02:00
|
|
|
if ((*last != '/') && depth) {
|
|
|
|
error("%s: ignoring relative alternate object store %s",
|
|
|
|
relative_base, last);
|
|
|
|
} else {
|
|
|
|
link_alt_odb_entry(last, cp - last,
|
|
|
|
relative_base, depth);
|
2005-12-05 07:48:43 +01:00
|
|
|
}
|
2005-08-15 02:25:57 +02:00
|
|
|
}
|
2005-08-17 03:22:05 +02:00
|
|
|
while (cp < ep && *cp == sep)
|
2005-08-15 02:25:57 +02:00
|
|
|
cp++;
|
|
|
|
last = cp;
|
2005-08-17 03:22:05 +02:00
|
|
|
}
|
2005-08-15 02:25:57 +02:00
|
|
|
}
|
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
static void read_info_alternates(const char * relative_base, int depth)
|
2005-08-15 02:25:57 +02:00
|
|
|
{
|
2005-08-17 03:22:05 +02:00
|
|
|
char *map;
|
2005-08-15 02:25:57 +02:00
|
|
|
struct stat st;
|
2006-05-07 20:19:21 +02:00
|
|
|
char path[PATH_MAX];
|
|
|
|
int fd;
|
2005-08-15 02:25:57 +02:00
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
sprintf(path, "%s/info/alternates", relative_base);
|
2005-08-15 02:25:57 +02:00
|
|
|
fd = open(path, O_RDONLY);
|
|
|
|
if (fd < 0)
|
|
|
|
return;
|
|
|
|
if (fstat(fd, &st) || (st.st_size == 0)) {
|
|
|
|
close(fd);
|
2005-06-28 23:56:57 +02:00
|
|
|
return;
|
2005-05-07 09:38:04 +02:00
|
|
|
}
|
2005-08-15 02:25:57 +02:00
|
|
|
map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
|
|
|
|
close(fd);
|
|
|
|
if (map == MAP_FAILED)
|
|
|
|
return;
|
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
link_alt_odb_entries(map, map + st.st_size, '\n', relative_base, depth);
|
|
|
|
|
2005-08-15 02:25:57 +02:00
|
|
|
munmap(map, st.st_size);
|
2005-05-07 09:38:04 +02:00
|
|
|
}
|
|
|
|
|
2006-05-07 20:19:21 +02:00
|
|
|
void prepare_alt_odb(void)
|
|
|
|
{
|
2006-06-28 11:04:39 +02:00
|
|
|
const char *alt;
|
2006-05-07 20:19:21 +02:00
|
|
|
|
|
|
|
alt = getenv(ALTERNATE_DB_ENVIRONMENT);
|
|
|
|
if (!alt) alt = "";
|
|
|
|
|
|
|
|
if (alt_odb_tail)
|
|
|
|
return;
|
|
|
|
alt_odb_tail = &alt_odb_list;
|
|
|
|
link_alt_odb_entries(alt, alt + strlen(alt), ':', NULL, 0);
|
|
|
|
|
|
|
|
read_info_alternates(get_object_directory(), 0);
|
|
|
|
}
|
|
|
|
|
2005-05-07 09:38:04 +02:00
|
|
|
static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
|
|
|
|
{
|
|
|
|
char *name = sha1_file_name(sha1);
|
2005-08-15 02:25:57 +02:00
|
|
|
struct alternate_object_database *alt;
|
2005-05-07 09:38:04 +02:00
|
|
|
|
|
|
|
if (!stat(name, st))
|
|
|
|
return name;
|
2005-06-28 23:56:57 +02:00
|
|
|
prepare_alt_odb();
|
2005-08-15 02:25:57 +02:00
|
|
|
for (alt = alt_odb_list; alt; alt = alt->next) {
|
|
|
|
name = alt->name;
|
2005-05-07 09:38:04 +02:00
|
|
|
fill_sha1_path(name, sha1);
|
2005-08-15 02:25:57 +02:00
|
|
|
if (!stat(alt->base, st))
|
|
|
|
return alt->base;
|
2005-05-07 09:38:04 +02:00
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-06-27 12:35:33 +02:00
|
|
|
#define PACK_MAX_SZ (1<<26)
|
|
|
|
static int pack_used_ctr;
|
|
|
|
static unsigned long pack_mapped;
|
2005-06-28 23:56:57 +02:00
|
|
|
struct packed_git *packed_git;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
|
|
|
static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
|
|
|
|
void **idx_map_)
|
|
|
|
{
|
|
|
|
void *idx_map;
|
|
|
|
unsigned int *index;
|
|
|
|
unsigned long idx_size;
|
|
|
|
int nr, i;
|
2006-01-19 05:26:14 +01:00
|
|
|
int fd = open(path, O_RDONLY);
|
2005-06-27 12:35:33 +02:00
|
|
|
struct stat st;
|
|
|
|
if (fd < 0)
|
|
|
|
return -1;
|
|
|
|
if (fstat(fd, &st)) {
|
|
|
|
close(fd);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
idx_size = st.st_size;
|
|
|
|
idx_map = mmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
|
|
|
|
close(fd);
|
|
|
|
if (idx_map == MAP_FAILED)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
index = idx_map;
|
2005-07-03 18:58:44 +02:00
|
|
|
*idx_map_ = idx_map;
|
|
|
|
*idx_size_ = idx_size;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
|
|
|
/* check index map */
|
2005-06-29 11:51:27 +02:00
|
|
|
if (idx_size < 4*256 + 20 + 20)
|
2005-06-27 12:35:33 +02:00
|
|
|
return error("index file too small");
|
|
|
|
nr = 0;
|
|
|
|
for (i = 0; i < 256; i++) {
|
|
|
|
unsigned int n = ntohl(index[i]);
|
|
|
|
if (n < nr)
|
|
|
|
return error("non-monotonic index");
|
|
|
|
nr = n;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Total size:
|
|
|
|
* - 256 index entries 4 bytes each
|
|
|
|
* - 24-byte entries * nr (20-byte sha1 + 4-byte offset)
|
|
|
|
* - 20-byte SHA1 of the packfile
|
|
|
|
* - 20-byte SHA1 file checksum
|
|
|
|
*/
|
|
|
|
if (idx_size != 4*256 + nr * 24 + 20 + 20)
|
|
|
|
return error("wrong index file size");
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-06-29 11:51:27 +02:00
|
|
|
static int unuse_one_packed_git(void)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
2005-06-29 11:51:27 +02:00
|
|
|
struct packed_git *p, *lru = NULL;
|
|
|
|
|
|
|
|
for (p = packed_git; p; p = p->next) {
|
|
|
|
if (p->pack_use_cnt || !p->pack_base)
|
|
|
|
continue;
|
|
|
|
if (!lru || p->pack_last_used < lru->pack_last_used)
|
|
|
|
lru = p;
|
|
|
|
}
|
|
|
|
if (!lru)
|
|
|
|
return 0;
|
|
|
|
munmap(lru->pack_base, lru->pack_size);
|
|
|
|
lru->pack_base = NULL;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
void unuse_packed_git(struct packed_git *p)
|
|
|
|
{
|
|
|
|
p->pack_use_cnt--;
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
|
2005-06-29 11:51:27 +02:00
|
|
|
int use_packed_git(struct packed_git *p)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
2005-08-01 02:53:44 +02:00
|
|
|
if (!p->pack_size) {
|
|
|
|
struct stat st;
|
2006-07-10 08:57:51 +02:00
|
|
|
/* We created the struct before we had the pack */
|
2005-08-01 02:53:44 +02:00
|
|
|
stat(p->pack_name, &st);
|
|
|
|
if (!S_ISREG(st.st_mode))
|
|
|
|
die("packfile %s not a regular file", p->pack_name);
|
|
|
|
p->pack_size = st.st_size;
|
|
|
|
}
|
2005-06-27 12:35:33 +02:00
|
|
|
if (!p->pack_base) {
|
|
|
|
int fd;
|
|
|
|
struct stat st;
|
|
|
|
void *map;
|
2006-08-21 09:10:02 +02:00
|
|
|
struct pack_header *hdr;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
|
|
|
pack_mapped += p->pack_size;
|
2005-06-29 11:51:27 +02:00
|
|
|
while (PACK_MAX_SZ < pack_mapped && unuse_one_packed_git())
|
|
|
|
; /* nothing */
|
2005-06-27 12:35:33 +02:00
|
|
|
fd = open(p->pack_name, O_RDONLY);
|
|
|
|
if (fd < 0)
|
2005-06-29 11:51:27 +02:00
|
|
|
die("packfile %s cannot be opened", p->pack_name);
|
2005-06-27 12:35:33 +02:00
|
|
|
if (fstat(fd, &st)) {
|
|
|
|
close(fd);
|
2005-06-29 11:51:27 +02:00
|
|
|
die("packfile %s cannot be opened", p->pack_name);
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
if (st.st_size != p->pack_size)
|
2005-06-29 11:51:27 +02:00
|
|
|
die("packfile %s size mismatch.", p->pack_name);
|
2005-06-27 12:35:33 +02:00
|
|
|
map = mmap(NULL, p->pack_size, PROT_READ, MAP_PRIVATE, fd, 0);
|
|
|
|
close(fd);
|
|
|
|
if (map == MAP_FAILED)
|
2005-06-29 11:51:27 +02:00
|
|
|
die("packfile %s cannot be mapped.", p->pack_name);
|
2005-06-27 12:35:33 +02:00
|
|
|
p->pack_base = map;
|
2005-06-29 11:51:27 +02:00
|
|
|
|
2006-08-21 09:10:02 +02:00
|
|
|
/* Check if we understand this pack file. If we don't we're
|
|
|
|
* likely too old to handle it.
|
|
|
|
*/
|
|
|
|
hdr = map;
|
|
|
|
if (hdr->hdr_signature != htonl(PACK_SIGNATURE))
|
|
|
|
die("packfile %s isn't actually a pack.", p->pack_name);
|
|
|
|
if (!pack_version_ok(hdr->hdr_version))
|
|
|
|
die("packfile %s is version %i and not supported"
|
|
|
|
" (try upgrading GIT to a newer version)",
|
|
|
|
p->pack_name, ntohl(hdr->hdr_version));
|
|
|
|
|
2005-06-29 11:51:27 +02:00
|
|
|
/* Check if the pack file matches with the index file.
|
|
|
|
* this is cheap.
|
|
|
|
*/
|
2006-08-17 20:54:57 +02:00
|
|
|
if (hashcmp((unsigned char *)(p->index_base) +
|
|
|
|
p->index_size - 40,
|
|
|
|
(unsigned char *)p->pack_base +
|
|
|
|
p->pack_size - 20)) {
|
2005-06-29 11:51:27 +02:00
|
|
|
die("packfile %s does not match index.", p->pack_name);
|
2005-08-01 02:53:44 +02:00
|
|
|
}
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
p->pack_last_used = pack_used_ctr++;
|
2005-06-29 11:51:27 +02:00
|
|
|
p->pack_use_cnt++;
|
2005-06-27 12:35:33 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-10-14 00:38:28 +02:00
|
|
|
struct packed_git *add_packed_git(char *path, int path_len, int local)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
|
|
|
struct stat st;
|
|
|
|
struct packed_git *p;
|
|
|
|
unsigned long idx_size;
|
|
|
|
void *idx_map;
|
2005-11-15 21:51:02 +01:00
|
|
|
unsigned char sha1[20];
|
2005-06-27 12:35:33 +02:00
|
|
|
|
|
|
|
if (check_packed_git_idx(path, &idx_size, &idx_map))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
/* do we have a corresponding .pack file? */
|
|
|
|
strcpy(path + path_len - 4, ".pack");
|
|
|
|
if (stat(path, &st) || !S_ISREG(st.st_mode)) {
|
|
|
|
munmap(idx_map, idx_size);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
/* ok, it looks sane as far as we can check without
|
|
|
|
* actually mapping the pack file.
|
|
|
|
*/
|
|
|
|
p = xmalloc(sizeof(*p) + path_len + 2);
|
|
|
|
strcpy(p->pack_name, path);
|
|
|
|
p->index_size = idx_size;
|
|
|
|
p->pack_size = st.st_size;
|
|
|
|
p->index_base = idx_map;
|
|
|
|
p->next = NULL;
|
2005-06-28 23:55:16 +02:00
|
|
|
p->pack_base = NULL;
|
2005-06-27 12:35:33 +02:00
|
|
|
p->pack_last_used = 0;
|
2005-06-29 11:51:27 +02:00
|
|
|
p->pack_use_cnt = 0;
|
2005-10-14 00:38:28 +02:00
|
|
|
p->pack_local = local;
|
2005-12-22 00:47:09 +01:00
|
|
|
if ((path_len > 44) && !get_sha1_hex(path + path_len - 44, sha1))
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(p->sha1, sha1);
|
2005-06-27 12:35:33 +02:00
|
|
|
return p;
|
|
|
|
}
|
|
|
|
|
2005-08-01 02:53:44 +02:00
|
|
|
struct packed_git *parse_pack_index(unsigned char *sha1)
|
2005-08-16 06:10:03 +02:00
|
|
|
{
|
|
|
|
char *path = sha1_pack_index_name(sha1);
|
|
|
|
return parse_pack_index_file(sha1, path);
|
|
|
|
}
|
|
|
|
|
2005-09-02 14:17:10 +02:00
|
|
|
struct packed_git *parse_pack_index_file(const unsigned char *sha1, char *idx_path)
|
2005-08-01 02:53:44 +02:00
|
|
|
{
|
|
|
|
struct packed_git *p;
|
|
|
|
unsigned long idx_size;
|
|
|
|
void *idx_map;
|
2005-08-16 06:10:03 +02:00
|
|
|
char *path;
|
2005-08-01 02:53:44 +02:00
|
|
|
|
2005-08-16 06:10:03 +02:00
|
|
|
if (check_packed_git_idx(idx_path, &idx_size, &idx_map))
|
2005-08-01 02:53:44 +02:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
path = sha1_pack_name(sha1);
|
|
|
|
|
|
|
|
p = xmalloc(sizeof(*p) + strlen(path) + 2);
|
|
|
|
strcpy(p->pack_name, path);
|
|
|
|
p->index_size = idx_size;
|
|
|
|
p->pack_size = 0;
|
|
|
|
p->index_base = idx_map;
|
|
|
|
p->next = NULL;
|
|
|
|
p->pack_base = NULL;
|
|
|
|
p->pack_last_used = 0;
|
|
|
|
p->pack_use_cnt = 0;
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(p->sha1, sha1);
|
2005-08-01 02:53:44 +02:00
|
|
|
return p;
|
|
|
|
}
|
|
|
|
|
|
|
|
void install_packed_git(struct packed_git *pack)
|
|
|
|
{
|
|
|
|
pack->next = packed_git;
|
|
|
|
packed_git = pack;
|
|
|
|
}
|
|
|
|
|
2005-10-14 00:38:28 +02:00
|
|
|
static void prepare_packed_git_one(char *objdir, int local)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
|
|
|
char path[PATH_MAX];
|
|
|
|
int len;
|
|
|
|
DIR *dir;
|
|
|
|
struct dirent *de;
|
|
|
|
|
|
|
|
sprintf(path, "%s/pack", objdir);
|
|
|
|
len = strlen(path);
|
|
|
|
dir = opendir(path);
|
2006-02-18 01:14:52 +01:00
|
|
|
if (!dir) {
|
2006-02-22 20:16:38 +01:00
|
|
|
if (errno != ENOENT)
|
2006-02-23 02:47:10 +01:00
|
|
|
error("unable to open object pack directory: %s: %s",
|
2006-02-22 20:16:38 +01:00
|
|
|
path, strerror(errno));
|
2005-06-27 12:35:33 +02:00
|
|
|
return;
|
2006-02-18 01:14:52 +01:00
|
|
|
}
|
2005-06-27 12:35:33 +02:00
|
|
|
path[len++] = '/';
|
|
|
|
while ((de = readdir(dir)) != NULL) {
|
|
|
|
int namelen = strlen(de->d_name);
|
|
|
|
struct packed_git *p;
|
|
|
|
|
2006-08-11 14:01:45 +02:00
|
|
|
if (!has_extension(de->d_name, ".idx"))
|
2005-06-27 12:35:33 +02:00
|
|
|
continue;
|
|
|
|
|
|
|
|
/* we have .idx. Is it a file we can map? */
|
|
|
|
strcpy(path + len, de->d_name);
|
2006-06-02 18:49:32 +02:00
|
|
|
for (p = packed_git; p; p = p->next) {
|
|
|
|
if (!memcmp(path, p->pack_name, len + namelen - 4))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (p)
|
|
|
|
continue;
|
2005-10-14 00:38:28 +02:00
|
|
|
p = add_packed_git(path, len + namelen, local);
|
2005-06-27 12:35:33 +02:00
|
|
|
if (!p)
|
|
|
|
continue;
|
|
|
|
p->next = packed_git;
|
|
|
|
packed_git = p;
|
|
|
|
}
|
2005-07-06 08:52:17 +02:00
|
|
|
closedir(dir);
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
|
2006-06-02 17:32:23 +02:00
|
|
|
static int prepare_packed_git_run_once = 0;
|
2005-06-28 23:56:57 +02:00
|
|
|
void prepare_packed_git(void)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
2005-08-15 02:25:57 +02:00
|
|
|
struct alternate_object_database *alt;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
2006-06-02 17:32:23 +02:00
|
|
|
if (prepare_packed_git_run_once)
|
2005-06-27 12:35:33 +02:00
|
|
|
return;
|
2005-10-14 00:38:28 +02:00
|
|
|
prepare_packed_git_one(get_object_directory(), 1);
|
2005-06-28 23:56:57 +02:00
|
|
|
prepare_alt_odb();
|
2005-08-15 02:25:57 +02:00
|
|
|
for (alt = alt_odb_list; alt; alt = alt->next) {
|
2005-12-05 07:48:43 +01:00
|
|
|
alt->name[-1] = 0;
|
2005-10-14 00:38:28 +02:00
|
|
|
prepare_packed_git_one(alt->base, 0);
|
2005-12-05 07:48:43 +01:00
|
|
|
alt->name[-1] = '/';
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
2006-06-02 17:32:23 +02:00
|
|
|
prepare_packed_git_run_once = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void reprepare_packed_git(void)
|
|
|
|
{
|
|
|
|
prepare_packed_git_run_once = 0;
|
|
|
|
prepare_packed_git();
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
|
2005-06-03 17:05:39 +02:00
|
|
|
int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
|
2005-04-18 22:04:43 +02:00
|
|
|
{
|
|
|
|
unsigned char real_sha1[20];
|
2006-10-15 14:02:18 +02:00
|
|
|
hash_sha1_file(map, size, type, real_sha1);
|
2006-08-17 20:54:57 +02:00
|
|
|
return hashcmp(sha1, real_sha1) ? -1 : 0;
|
2005-04-18 22:04:43 +02:00
|
|
|
}
|
|
|
|
|
2006-07-18 00:04:47 +02:00
|
|
|
void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
|
2005-04-18 22:04:43 +02:00
|
|
|
{
|
|
|
|
struct stat st;
|
|
|
|
void *map;
|
2005-04-23 20:09:32 +02:00
|
|
|
int fd;
|
2005-05-07 09:38:04 +02:00
|
|
|
char *filename = find_sha1_file(sha1, &st);
|
|
|
|
|
|
|
|
if (!filename) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2005-04-18 22:04:43 +02:00
|
|
|
|
2005-04-23 20:09:32 +02:00
|
|
|
fd = open(filename, O_RDONLY | sha1_file_open_flag);
|
2005-04-18 22:04:43 +02:00
|
|
|
if (fd < 0) {
|
2005-04-23 20:09:32 +02:00
|
|
|
/* See if it works without O_NOATIME */
|
|
|
|
switch (sha1_file_open_flag) {
|
|
|
|
default:
|
|
|
|
fd = open(filename, O_RDONLY);
|
|
|
|
if (fd >= 0)
|
|
|
|
break;
|
|
|
|
/* Fallthrough */
|
|
|
|
case 0:
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-06-27 12:35:33 +02:00
|
|
|
/* If it failed once, it will probably fail again.
|
|
|
|
* Stop using O_NOATIME
|
|
|
|
*/
|
2005-04-23 20:09:32 +02:00
|
|
|
sha1_file_open_flag = 0;
|
2005-04-18 22:04:43 +02:00
|
|
|
}
|
|
|
|
map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
|
|
|
|
close(fd);
|
2005-07-29 16:49:14 +02:00
|
|
|
if (map == MAP_FAILED)
|
2005-04-18 22:04:43 +02:00
|
|
|
return NULL;
|
|
|
|
*size = st.st_size;
|
|
|
|
return map;
|
|
|
|
}
|
|
|
|
|
2006-07-18 00:04:47 +02:00
|
|
|
int legacy_loose_object(unsigned char *map)
|
|
|
|
{
|
|
|
|
unsigned int word;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Is it a zlib-compressed buffer? If so, the first byte
|
|
|
|
* must be 0x78 (15-bit window size, deflated), and the
|
|
|
|
* first 16-bit word is evenly divisible by 31
|
|
|
|
*/
|
|
|
|
word = (map[0] << 8) + map[1];
|
|
|
|
if (map[0] == 0x78 && !(word % 31))
|
|
|
|
return 1;
|
|
|
|
else
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-09-04 06:09:18 +02:00
|
|
|
unsigned long unpack_object_header_gently(const unsigned char *buf, unsigned long len, enum object_type *type, unsigned long *sizep)
|
2005-06-02 02:54:59 +02:00
|
|
|
{
|
2006-09-02 00:17:01 +02:00
|
|
|
unsigned shift;
|
2006-07-11 21:48:08 +02:00
|
|
|
unsigned char c;
|
|
|
|
unsigned long size;
|
2006-09-02 00:17:01 +02:00
|
|
|
unsigned long used = 0;
|
|
|
|
|
|
|
|
c = buf[used++];
|
|
|
|
*type = (c >> 4) & 7;
|
|
|
|
size = c & 15;
|
|
|
|
shift = 4;
|
|
|
|
while (c & 0x80) {
|
|
|
|
if (len <= used)
|
|
|
|
return 0;
|
|
|
|
if (sizeof(long) * 8 <= shift)
|
|
|
|
return 0;
|
|
|
|
c = buf[used++];
|
|
|
|
size += (c & 0x7f) << shift;
|
|
|
|
shift += 7;
|
|
|
|
}
|
|
|
|
*sizep = size;
|
|
|
|
return used;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int unpack_sha1_header(z_stream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz)
|
|
|
|
{
|
|
|
|
unsigned long size, used;
|
|
|
|
static const char valid_loose_object_type[8] = {
|
|
|
|
0, /* OBJ_EXT */
|
|
|
|
1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */
|
|
|
|
0, /* "delta" and others are invalid in a loose object */
|
2006-07-11 21:48:08 +02:00
|
|
|
};
|
2006-09-02 00:17:01 +02:00
|
|
|
enum object_type type;
|
2006-07-11 21:48:08 +02:00
|
|
|
|
2005-06-02 02:54:59 +02:00
|
|
|
/* Get the data stream */
|
|
|
|
memset(stream, 0, sizeof(*stream));
|
|
|
|
stream->next_in = map;
|
|
|
|
stream->avail_in = mapsize;
|
|
|
|
stream->next_out = buffer;
|
2006-07-11 21:48:08 +02:00
|
|
|
stream->avail_out = bufsiz;
|
|
|
|
|
2006-07-18 00:04:47 +02:00
|
|
|
if (legacy_loose_object(map)) {
|
2006-07-11 21:48:08 +02:00
|
|
|
inflateInit(stream);
|
|
|
|
return inflate(stream, 0);
|
|
|
|
}
|
|
|
|
|
2006-09-02 00:17:01 +02:00
|
|
|
used = unpack_object_header_gently(map, mapsize, &type, &size);
|
|
|
|
if (!used || !valid_loose_object_type[type])
|
2006-07-11 21:48:08 +02:00
|
|
|
return -1;
|
2006-09-02 00:17:01 +02:00
|
|
|
map += used;
|
|
|
|
mapsize -= used;
|
2006-07-11 21:48:08 +02:00
|
|
|
|
|
|
|
/* Set up the stream for the rest.. */
|
|
|
|
stream->next_in = map;
|
|
|
|
stream->avail_in = mapsize;
|
2005-06-02 02:54:59 +02:00
|
|
|
inflateInit(stream);
|
2006-07-11 21:48:08 +02:00
|
|
|
|
|
|
|
/* And generate the fake traditional header */
|
2006-09-02 00:17:01 +02:00
|
|
|
stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu",
|
|
|
|
type_names[type], size);
|
2006-07-11 21:48:08 +02:00
|
|
|
return 0;
|
2005-06-02 02:54:59 +02:00
|
|
|
}
|
|
|
|
|
2005-07-03 19:10:45 +02:00
|
|
|
static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size)
|
2005-06-02 16:57:25 +02:00
|
|
|
{
|
|
|
|
int bytes = strlen(buffer) + 1;
|
2005-06-21 22:04:33 +02:00
|
|
|
unsigned char *buf = xmalloc(1+size);
|
2006-07-11 21:48:08 +02:00
|
|
|
unsigned long n;
|
2005-06-02 16:57:25 +02:00
|
|
|
|
2006-07-11 21:48:08 +02:00
|
|
|
n = stream->total_out - bytes;
|
|
|
|
if (n > size)
|
|
|
|
n = size;
|
|
|
|
memcpy(buf, (char *) buffer + bytes, n);
|
|
|
|
bytes = n;
|
2005-06-02 16:57:25 +02:00
|
|
|
if (bytes < size) {
|
|
|
|
stream->next_out = buf + bytes;
|
|
|
|
stream->avail_out = size - bytes;
|
|
|
|
while (inflate(stream, Z_FINISH) == Z_OK)
|
|
|
|
/* nothing */;
|
|
|
|
}
|
|
|
|
buf[size] = 0;
|
|
|
|
inflateEnd(stream);
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We used to just use "sscanf()", but that's actually way
|
|
|
|
* too permissive for what we want to check. So do an anal
|
|
|
|
* object header parse by hand.
|
|
|
|
*/
|
2006-07-11 19:10:29 +02:00
|
|
|
static int parse_sha1_header(char *hdr, char *type, unsigned long *sizep)
|
2005-06-02 16:57:25 +02:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
unsigned long size;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The type can be at most ten bytes (including the
|
|
|
|
* terminating '\0' that we add), and is followed by
|
|
|
|
* a space.
|
|
|
|
*/
|
|
|
|
i = 10;
|
|
|
|
for (;;) {
|
|
|
|
char c = *hdr++;
|
|
|
|
if (c == ' ')
|
|
|
|
break;
|
|
|
|
if (!--i)
|
|
|
|
return -1;
|
|
|
|
*type++ = c;
|
|
|
|
}
|
|
|
|
*type = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The length must follow immediately, and be in canonical
|
|
|
|
* decimal format (ie "010" is not valid).
|
|
|
|
*/
|
|
|
|
size = *hdr++ - '0';
|
|
|
|
if (size > 9)
|
|
|
|
return -1;
|
|
|
|
if (size) {
|
|
|
|
for (;;) {
|
|
|
|
unsigned long c = *hdr - '0';
|
|
|
|
if (c > 9)
|
|
|
|
break;
|
|
|
|
hdr++;
|
|
|
|
size = size * 10 + c;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
*sizep = size;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The length must be followed by a zero byte
|
|
|
|
*/
|
|
|
|
return *hdr ? -1 : 0;
|
|
|
|
}
|
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
|
|
|
|
{
|
2005-06-02 16:57:25 +02:00
|
|
|
int ret;
|
2005-04-18 22:04:43 +02:00
|
|
|
z_stream stream;
|
2005-06-02 16:57:25 +02:00
|
|
|
char hdr[8192];
|
2005-04-18 22:04:43 +02:00
|
|
|
|
2005-06-02 16:57:25 +02:00
|
|
|
ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
|
|
|
|
if (ret < Z_OK || parse_sha1_header(hdr, type, size) < 0)
|
2005-04-18 22:04:43 +02:00
|
|
|
return NULL;
|
|
|
|
|
2005-06-02 16:57:25 +02:00
|
|
|
return unpack_sha1_rest(&stream, hdr, *size);
|
2005-04-18 22:04:43 +02:00
|
|
|
}
|
|
|
|
|
2005-07-01 02:15:39 +02:00
|
|
|
/* forward declaration for a mutually recursive function */
|
2006-09-21 06:05:37 +02:00
|
|
|
static int packed_object_info(struct packed_git *p, unsigned long offset,
|
2005-07-01 02:15:39 +02:00
|
|
|
char *type, unsigned long *sizep);
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
static int packed_delta_info(struct packed_git *p,
|
|
|
|
unsigned long offset,
|
2005-06-28 08:58:08 +02:00
|
|
|
char *type,
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long *sizep)
|
2005-06-28 08:58:08 +02:00
|
|
|
{
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long base_offset;
|
|
|
|
unsigned char *base_sha1 = (unsigned char *) p->pack_base + offset;
|
2005-07-01 02:15:39 +02:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
if (p->pack_size < offset + 20)
|
2005-06-28 08:58:08 +02:00
|
|
|
die("truncated pack file");
|
2005-07-01 02:15:39 +02:00
|
|
|
/* The base entry _must_ be in the same pack */
|
2006-09-21 06:05:37 +02:00
|
|
|
base_offset = find_pack_entry_one(base_sha1, p);
|
|
|
|
if (!base_offset)
|
2005-07-01 02:15:39 +02:00
|
|
|
die("failed to find delta-pack base object %s",
|
|
|
|
sha1_to_hex(base_sha1));
|
2006-09-21 06:05:37 +02:00
|
|
|
offset += 20;
|
2005-07-01 02:15:39 +02:00
|
|
|
|
2005-07-01 02:13:07 +02:00
|
|
|
/* We choose to only get the type of the base object and
|
|
|
|
* ignore potentially corrupt pack file that expects the delta
|
|
|
|
* based on a base with a wrong size. This saves tons of
|
|
|
|
* inflate() calls.
|
|
|
|
*/
|
2006-09-21 06:05:37 +02:00
|
|
|
if (packed_object_info(p, base_offset, type, NULL))
|
2005-07-01 02:13:07 +02:00
|
|
|
die("cannot get info for delta-pack base");
|
2005-06-28 08:58:08 +02:00
|
|
|
|
2005-07-01 02:13:07 +02:00
|
|
|
if (sizep) {
|
|
|
|
const unsigned char *data;
|
|
|
|
unsigned char delta_head[64];
|
|
|
|
unsigned long result_size;
|
|
|
|
z_stream stream;
|
|
|
|
int st;
|
|
|
|
|
|
|
|
memset(&stream, 0, sizeof(stream));
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
stream.next_in = (unsigned char *) p->pack_base + offset;
|
|
|
|
stream.avail_in = p->pack_size - offset;
|
2005-07-01 02:13:07 +02:00
|
|
|
stream.next_out = delta_head;
|
|
|
|
stream.avail_out = sizeof(delta_head);
|
|
|
|
|
|
|
|
inflateInit(&stream);
|
|
|
|
st = inflate(&stream, Z_FINISH);
|
|
|
|
inflateEnd(&stream);
|
|
|
|
if ((st != Z_STREAM_END) &&
|
|
|
|
stream.total_out != sizeof(delta_head))
|
|
|
|
die("delta data unpack-initial failed");
|
|
|
|
|
|
|
|
/* Examine the initial part of the delta to figure out
|
|
|
|
* the result size.
|
|
|
|
*/
|
|
|
|
data = delta_head;
|
2006-04-07 21:26:10 +02:00
|
|
|
|
|
|
|
/* ignore base size */
|
|
|
|
get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
|
2005-06-28 08:58:08 +02:00
|
|
|
|
2005-07-01 02:13:07 +02:00
|
|
|
/* Read the result size */
|
2006-04-07 21:26:10 +02:00
|
|
|
result_size = get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
|
2005-07-01 02:13:07 +02:00
|
|
|
*sizep = result_size;
|
|
|
|
}
|
2005-06-28 08:58:08 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-06-28 23:21:02 +02:00
|
|
|
static unsigned long unpack_object_header(struct packed_git *p, unsigned long offset,
|
|
|
|
enum object_type *type, unsigned long *sizep)
|
|
|
|
{
|
2006-09-02 00:17:01 +02:00
|
|
|
unsigned long used;
|
2005-06-28 23:21:02 +02:00
|
|
|
|
2006-09-02 00:17:01 +02:00
|
|
|
if (p->pack_size <= offset)
|
2005-06-28 23:21:02 +02:00
|
|
|
die("object offset outside of pack file");
|
2006-09-02 00:17:01 +02:00
|
|
|
|
|
|
|
used = unpack_object_header_gently((unsigned char *)p->pack_base +
|
|
|
|
offset,
|
|
|
|
p->pack_size - offset, type, sizep);
|
|
|
|
if (!used)
|
|
|
|
die("object offset outside of pack file");
|
|
|
|
|
|
|
|
return offset + used;
|
2005-06-28 23:21:02 +02:00
|
|
|
}
|
|
|
|
|
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 02:34:29 +01:00
|
|
|
int check_reuse_pack_delta(struct packed_git *p, unsigned long offset,
|
|
|
|
unsigned char *base, unsigned long *sizep,
|
|
|
|
enum object_type *kindp)
|
|
|
|
{
|
|
|
|
unsigned long ptr;
|
|
|
|
int status = -1;
|
|
|
|
|
|
|
|
use_packed_git(p);
|
|
|
|
ptr = offset;
|
|
|
|
ptr = unpack_object_header(p, ptr, kindp, sizep);
|
|
|
|
if (*kindp != OBJ_DELTA)
|
|
|
|
goto done;
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(base, (unsigned char *) p->pack_base + ptr);
|
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 02:34:29 +01:00
|
|
|
status = 0;
|
|
|
|
done:
|
|
|
|
unuse_packed_git(p);
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
void packed_object_info_detail(struct packed_git *p,
|
|
|
|
unsigned long offset,
|
2005-07-01 02:17:20 +02:00
|
|
|
char *type,
|
|
|
|
unsigned long *size,
|
|
|
|
unsigned long *store_size,
|
2006-02-15 21:47:43 +01:00
|
|
|
unsigned int *delta_chain_length,
|
2005-07-01 02:17:20 +02:00
|
|
|
unsigned char *base_sha1)
|
|
|
|
{
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long val;
|
|
|
|
unsigned char *next_sha1;
|
2005-07-01 02:17:20 +02:00
|
|
|
enum object_type kind;
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
*delta_chain_length = 0;
|
|
|
|
offset = unpack_object_header(p, offset, &kind, size);
|
|
|
|
|
|
|
|
for (;;) {
|
|
|
|
switch (kind) {
|
|
|
|
default:
|
|
|
|
die("corrupted pack file %s containing object of kind %d",
|
|
|
|
p->pack_name, kind);
|
|
|
|
case OBJ_COMMIT:
|
|
|
|
case OBJ_TREE:
|
|
|
|
case OBJ_BLOB:
|
|
|
|
case OBJ_TAG:
|
|
|
|
strcpy(type, type_names[kind]);
|
|
|
|
*store_size = 0; /* notyet */
|
|
|
|
return;
|
|
|
|
case OBJ_DELTA:
|
|
|
|
if (p->pack_size <= offset + 20)
|
|
|
|
die("pack file %s records an incomplete delta base",
|
|
|
|
p->pack_name);
|
|
|
|
next_sha1 = (unsigned char *) p->pack_base + offset;
|
|
|
|
if (*delta_chain_length == 0)
|
|
|
|
hashcpy(base_sha1, next_sha1);
|
|
|
|
offset = find_pack_entry_one(next_sha1, p);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
offset = unpack_object_header(p, offset, &kind, &val);
|
|
|
|
(*delta_chain_length)++;
|
2005-07-01 02:17:20 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
static int packed_object_info(struct packed_git *p, unsigned long offset,
|
2005-06-27 12:35:33 +02:00
|
|
|
char *type, unsigned long *sizep)
|
|
|
|
{
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long size;
|
2005-06-28 23:21:02 +02:00
|
|
|
enum object_type kind;
|
2005-06-28 08:58:08 +02:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
offset = unpack_object_header(p, offset, &kind, &size);
|
2005-06-28 08:58:08 +02:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
if (kind == OBJ_DELTA)
|
|
|
|
return packed_delta_info(p, offset, type, sizep);
|
2005-06-28 23:21:02 +02:00
|
|
|
|
|
|
|
switch (kind) {
|
|
|
|
case OBJ_COMMIT:
|
|
|
|
case OBJ_TREE:
|
|
|
|
case OBJ_BLOB:
|
|
|
|
case OBJ_TAG:
|
2006-08-26 10:11:36 +02:00
|
|
|
strcpy(type, type_names[kind]);
|
2005-06-28 18:58:23 +02:00
|
|
|
break;
|
2005-06-27 12:35:33 +02:00
|
|
|
default:
|
2005-09-30 09:09:04 +02:00
|
|
|
die("corrupted pack file %s containing object of kind %d",
|
|
|
|
p->pack_name, kind);
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
2005-07-01 02:13:07 +02:00
|
|
|
if (sizep)
|
|
|
|
*sizep = size;
|
2005-06-27 12:35:33 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-08-26 10:12:27 +02:00
|
|
|
static void *unpack_compressed_entry(struct packed_git *p,
|
|
|
|
unsigned long offset,
|
|
|
|
unsigned long size)
|
2006-08-26 10:10:43 +02:00
|
|
|
{
|
|
|
|
int st;
|
|
|
|
z_stream stream;
|
|
|
|
unsigned char *buffer;
|
|
|
|
|
|
|
|
buffer = xmalloc(size + 1);
|
|
|
|
buffer[size] = 0;
|
|
|
|
memset(&stream, 0, sizeof(stream));
|
2006-08-26 10:12:27 +02:00
|
|
|
stream.next_in = (unsigned char*)p->pack_base + offset;
|
|
|
|
stream.avail_in = p->pack_size - offset;
|
2006-08-26 10:10:43 +02:00
|
|
|
stream.next_out = buffer;
|
|
|
|
stream.avail_out = size;
|
|
|
|
|
|
|
|
inflateInit(&stream);
|
|
|
|
st = inflate(&stream, Z_FINISH);
|
|
|
|
inflateEnd(&stream);
|
|
|
|
if ((st != Z_STREAM_END) || stream.total_out != size) {
|
|
|
|
free(buffer);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return buffer;
|
|
|
|
}
|
|
|
|
|
2006-08-26 10:12:27 +02:00
|
|
|
static void *unpack_delta_entry(struct packed_git *p,
|
|
|
|
unsigned long offset,
|
2005-06-27 12:35:33 +02:00
|
|
|
unsigned long delta_size,
|
|
|
|
char *type,
|
2006-08-26 10:12:27 +02:00
|
|
|
unsigned long *sizep)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
2006-08-26 10:11:02 +02:00
|
|
|
void *delta_data, *result, *base;
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long result_size, base_size, base_offset;
|
|
|
|
unsigned char *base_sha1;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
if (p->pack_size < offset + 20)
|
2005-06-27 12:35:33 +02:00
|
|
|
die("truncated pack file");
|
2006-03-19 22:43:42 +01:00
|
|
|
/* The base entry _must_ be in the same pack */
|
2006-08-26 10:12:27 +02:00
|
|
|
base_sha1 = (unsigned char*)p->pack_base + offset;
|
2006-09-21 06:05:37 +02:00
|
|
|
base_offset = find_pack_entry_one(base_sha1, p);
|
|
|
|
if (!base_offset)
|
2006-03-19 22:43:42 +01:00
|
|
|
die("failed to find delta-pack base object %s",
|
|
|
|
sha1_to_hex(base_sha1));
|
2006-09-21 06:05:37 +02:00
|
|
|
offset += 20;
|
|
|
|
|
|
|
|
base = unpack_entry_gently(p, base_offset, type, &base_size);
|
2006-03-19 22:43:42 +01:00
|
|
|
if (!base)
|
2006-09-21 06:05:37 +02:00
|
|
|
die("failed to read delta base object at %lu from %s",
|
|
|
|
base_offset, p->pack_name);
|
2006-03-19 22:43:42 +01:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
delta_data = unpack_compressed_entry(p, offset, delta_size);
|
2005-06-27 12:35:33 +02:00
|
|
|
result = patch_delta(base, base_size,
|
|
|
|
delta_data, delta_size,
|
|
|
|
&result_size);
|
|
|
|
if (!result)
|
|
|
|
die("failed to apply delta");
|
|
|
|
free(delta_data);
|
|
|
|
free(base);
|
|
|
|
*sizep = result_size;
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *unpack_entry(struct pack_entry *entry,
|
|
|
|
char *type, unsigned long *sizep)
|
|
|
|
{
|
|
|
|
struct packed_git *p = entry->p;
|
2005-06-29 11:51:27 +02:00
|
|
|
void *retval;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
|
|
|
if (use_packed_git(p))
|
|
|
|
die("cannot map packed file");
|
2006-09-21 06:05:37 +02:00
|
|
|
retval = unpack_entry_gently(p, entry->offset, type, sizep);
|
2005-07-01 02:15:39 +02:00
|
|
|
unuse_packed_git(p);
|
|
|
|
if (!retval)
|
2005-09-30 09:09:04 +02:00
|
|
|
die("corrupted pack file %s", p->pack_name);
|
2005-07-01 02:15:39 +02:00
|
|
|
return retval;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* The caller is responsible for use_packed_git()/unuse_packed_git() pair */
|
2006-09-21 06:05:37 +02:00
|
|
|
void *unpack_entry_gently(struct packed_git *p, unsigned long offset,
|
2005-07-01 02:15:39 +02:00
|
|
|
char *type, unsigned long *sizep)
|
|
|
|
{
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long size;
|
2005-07-01 02:15:39 +02:00
|
|
|
enum object_type kind;
|
2005-06-27 12:35:33 +02:00
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
offset = unpack_object_header(p, offset, &kind, &size);
|
2005-06-28 23:21:02 +02:00
|
|
|
switch (kind) {
|
|
|
|
case OBJ_DELTA:
|
2006-08-26 10:12:27 +02:00
|
|
|
return unpack_delta_entry(p, offset, size, type, sizep);
|
2005-06-28 23:21:02 +02:00
|
|
|
case OBJ_COMMIT:
|
|
|
|
case OBJ_TREE:
|
|
|
|
case OBJ_BLOB:
|
|
|
|
case OBJ_TAG:
|
2006-08-26 10:11:36 +02:00
|
|
|
strcpy(type, type_names[kind]);
|
|
|
|
*sizep = size;
|
2006-08-26 10:12:27 +02:00
|
|
|
return unpack_compressed_entry(p, offset, size);
|
2005-06-27 12:35:33 +02:00
|
|
|
default:
|
2005-07-01 02:15:39 +02:00
|
|
|
return NULL;
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-06-28 23:56:57 +02:00
|
|
|
int num_packed_objects(const struct packed_git *p)
|
|
|
|
{
|
2005-06-29 11:51:27 +02:00
|
|
|
/* See check_packed_git_idx() */
|
2005-06-28 23:56:57 +02:00
|
|
|
return (p->index_size - 20 - 20 - 4*256) / 24;
|
|
|
|
}
|
|
|
|
|
|
|
|
int nth_packed_object_sha1(const struct packed_git *p, int n,
|
|
|
|
unsigned char* sha1)
|
|
|
|
{
|
|
|
|
void *index = p->index_base + 256;
|
|
|
|
if (n < 0 || num_packed_objects(p) <= n)
|
|
|
|
return -1;
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(sha1, (unsigned char *) index + (24 * n) + 4);
|
2005-06-28 23:56:57 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long find_pack_entry_one(const unsigned char *sha1,
|
|
|
|
struct packed_git *p)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
2005-07-03 18:58:44 +02:00
|
|
|
unsigned int *level1_ofs = p->index_base;
|
2005-06-27 12:35:33 +02:00
|
|
|
int hi = ntohl(level1_ofs[*sha1]);
|
|
|
|
int lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1]));
|
|
|
|
void *index = p->index_base + 256;
|
|
|
|
|
|
|
|
do {
|
|
|
|
int mi = (lo + hi) / 2;
|
2006-08-17 20:54:57 +02:00
|
|
|
int cmp = hashcmp((unsigned char *)index + (24 * mi) + 4, sha1);
|
2006-09-21 06:05:37 +02:00
|
|
|
if (!cmp)
|
|
|
|
return ntohl(*((unsigned int *) ((char *) index + (24 * mi))));
|
2005-06-27 12:35:33 +02:00
|
|
|
if (cmp > 0)
|
|
|
|
hi = mi;
|
|
|
|
else
|
|
|
|
lo = mi+1;
|
|
|
|
} while (lo < hi);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-09-06 11:12:09 +02:00
|
|
|
static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed)
|
2005-06-27 12:35:33 +02:00
|
|
|
{
|
|
|
|
struct packed_git *p;
|
2006-09-21 06:05:37 +02:00
|
|
|
unsigned long offset;
|
|
|
|
|
2005-06-27 12:35:33 +02:00
|
|
|
prepare_packed_git();
|
|
|
|
|
|
|
|
for (p = packed_git; p; p = p->next) {
|
2006-09-06 11:12:09 +02:00
|
|
|
if (ignore_packed) {
|
|
|
|
const char **ig;
|
|
|
|
for (ig = ignore_packed; *ig; ig++)
|
|
|
|
if (!strcmp(p->pack_name, *ig))
|
|
|
|
break;
|
|
|
|
if (*ig)
|
|
|
|
continue;
|
|
|
|
}
|
2006-09-21 06:05:37 +02:00
|
|
|
offset = find_pack_entry_one(sha1, p);
|
|
|
|
if (offset) {
|
|
|
|
e->offset = offset;
|
|
|
|
e->p = p;
|
|
|
|
hashcpy(e->sha1, sha1);
|
2005-06-27 12:35:33 +02:00
|
|
|
return 1;
|
2006-09-21 06:05:37 +02:00
|
|
|
}
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-08-01 02:53:44 +02:00
|
|
|
struct packed_git *find_sha1_pack(const unsigned char *sha1,
|
|
|
|
struct packed_git *packs)
|
|
|
|
{
|
|
|
|
struct packed_git *p;
|
|
|
|
|
|
|
|
for (p = packs; p; p = p->next) {
|
2006-09-21 06:05:37 +02:00
|
|
|
if (find_pack_entry_one(sha1, p))
|
2005-08-01 02:53:44 +02:00
|
|
|
return p;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2005-06-27 12:34:06 +02:00
|
|
|
int sha1_object_info(const unsigned char *sha1, char *type, unsigned long *sizep)
|
2005-06-03 00:20:54 +02:00
|
|
|
{
|
2005-06-27 12:34:06 +02:00
|
|
|
int status;
|
2005-06-03 00:20:54 +02:00
|
|
|
unsigned long mapsize, size;
|
|
|
|
void *map;
|
|
|
|
z_stream stream;
|
2005-06-27 12:34:06 +02:00
|
|
|
char hdr[128];
|
2005-06-03 00:20:54 +02:00
|
|
|
|
2006-07-18 00:04:47 +02:00
|
|
|
map = map_sha1_file(sha1, &mapsize);
|
2005-06-27 12:35:33 +02:00
|
|
|
if (!map) {
|
|
|
|
struct pack_entry e;
|
|
|
|
|
2006-09-21 06:05:37 +02:00
|
|
|
if (!find_pack_entry(sha1, &e, NULL)) {
|
|
|
|
reprepare_packed_git();
|
|
|
|
if (!find_pack_entry(sha1, &e, NULL))
|
|
|
|
return error("unable to find %s", sha1_to_hex(sha1));
|
|
|
|
}
|
|
|
|
if (use_packed_git(e.p))
|
|
|
|
die("cannot map packed file");
|
|
|
|
status = packed_object_info(e.p, e.offset, type, sizep);
|
|
|
|
unuse_packed_git(e.p);
|
|
|
|
return status;
|
2005-06-27 12:35:33 +02:00
|
|
|
}
|
2005-06-27 12:34:06 +02:00
|
|
|
if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
|
|
|
|
status = error("unable to unpack %s header",
|
|
|
|
sha1_to_hex(sha1));
|
|
|
|
if (parse_sha1_header(hdr, type, &size) < 0)
|
|
|
|
status = error("unable to parse %s header", sha1_to_hex(sha1));
|
2005-06-27 12:33:33 +02:00
|
|
|
else {
|
2005-06-03 00:20:54 +02:00
|
|
|
status = 0;
|
2005-07-01 02:13:07 +02:00
|
|
|
if (sizep)
|
|
|
|
*sizep = size;
|
2005-06-03 00:20:54 +02:00
|
|
|
}
|
|
|
|
inflateEnd(&stream);
|
|
|
|
munmap(map, mapsize);
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
2005-06-27 12:35:33 +02:00
|
|
|
static void *read_packed_sha1(const unsigned char *sha1, char *type, unsigned long *size)
|
|
|
|
{
|
|
|
|
struct pack_entry e;
|
|
|
|
|
2006-09-06 11:12:09 +02:00
|
|
|
if (!find_pack_entry(sha1, &e, NULL)) {
|
2005-06-27 12:35:33 +02:00
|
|
|
error("cannot read sha1_file for %s", sha1_to_hex(sha1));
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
return unpack_entry(&e, type, size);
|
|
|
|
}
|
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
|
|
|
|
{
|
|
|
|
unsigned long mapsize;
|
|
|
|
void *map, *buf;
|
2005-07-11 09:00:55 +02:00
|
|
|
struct pack_entry e;
|
2005-04-18 22:04:43 +02:00
|
|
|
|
2006-09-06 11:12:09 +02:00
|
|
|
if (find_pack_entry(sha1, &e, NULL))
|
2005-07-11 09:00:55 +02:00
|
|
|
return read_packed_sha1(sha1, type, size);
|
2006-07-18 00:04:47 +02:00
|
|
|
map = map_sha1_file(sha1, &mapsize);
|
2005-04-18 22:04:43 +02:00
|
|
|
if (map) {
|
|
|
|
buf = unpack_sha1_file(map, mapsize, type, size);
|
|
|
|
munmap(map, mapsize);
|
|
|
|
return buf;
|
|
|
|
}
|
2006-06-02 17:32:23 +02:00
|
|
|
reprepare_packed_git();
|
2006-09-06 11:12:09 +02:00
|
|
|
if (find_pack_entry(sha1, &e, NULL))
|
2006-06-02 17:32:23 +02:00
|
|
|
return read_packed_sha1(sha1, type, size);
|
2005-07-11 09:00:55 +02:00
|
|
|
return NULL;
|
2005-04-18 22:04:43 +02:00
|
|
|
}
|
|
|
|
|
2005-04-29 01:42:27 +02:00
|
|
|
void *read_object_with_reference(const unsigned char *sha1,
|
2005-05-18 14:14:09 +02:00
|
|
|
const char *required_type,
|
2005-04-29 01:42:27 +02:00
|
|
|
unsigned long *size,
|
|
|
|
unsigned char *actual_sha1_return)
|
2005-04-21 03:06:49 +02:00
|
|
|
{
|
|
|
|
char type[20];
|
|
|
|
void *buffer;
|
|
|
|
unsigned long isize;
|
2005-04-29 01:42:27 +02:00
|
|
|
unsigned char actual_sha1[20];
|
2005-04-21 03:06:49 +02:00
|
|
|
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(actual_sha1, sha1);
|
2005-04-29 01:42:27 +02:00
|
|
|
while (1) {
|
|
|
|
int ref_length = -1;
|
|
|
|
const char *ref_type = NULL;
|
2005-04-21 03:06:49 +02:00
|
|
|
|
2005-04-29 01:42:27 +02:00
|
|
|
buffer = read_sha1_file(actual_sha1, type, &isize);
|
|
|
|
if (!buffer)
|
|
|
|
return NULL;
|
|
|
|
if (!strcmp(type, required_type)) {
|
|
|
|
*size = isize;
|
|
|
|
if (actual_sha1_return)
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(actual_sha1_return, actual_sha1);
|
2005-04-29 01:42:27 +02:00
|
|
|
return buffer;
|
|
|
|
}
|
|
|
|
/* Handle references */
|
2006-04-02 14:44:09 +02:00
|
|
|
else if (!strcmp(type, commit_type))
|
2005-04-29 01:42:27 +02:00
|
|
|
ref_type = "tree ";
|
2006-04-02 14:44:09 +02:00
|
|
|
else if (!strcmp(type, tag_type))
|
2005-04-29 01:42:27 +02:00
|
|
|
ref_type = "object ";
|
|
|
|
else {
|
|
|
|
free(buffer);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
ref_length = strlen(ref_type);
|
2005-04-21 03:06:49 +02:00
|
|
|
|
2005-04-29 01:42:27 +02:00
|
|
|
if (memcmp(buffer, ref_type, ref_length) ||
|
2006-06-18 17:18:09 +02:00
|
|
|
get_sha1_hex((char *) buffer + ref_length, actual_sha1)) {
|
2005-04-29 01:42:27 +02:00
|
|
|
free(buffer);
|
|
|
|
return NULL;
|
|
|
|
}
|
2005-08-08 20:44:43 +02:00
|
|
|
free(buffer);
|
2005-04-29 01:42:27 +02:00
|
|
|
/* Now we have the ID of the referred-to object in
|
|
|
|
* actual_sha1. Check again. */
|
2005-04-21 03:06:49 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-10-15 14:02:03 +02:00
|
|
|
static void write_sha1_file_prepare(void *buf, unsigned long len,
|
|
|
|
const char *type, unsigned char *sha1,
|
|
|
|
unsigned char *hdr, int *hdrlen)
|
2005-06-28 04:03:13 +02:00
|
|
|
{
|
|
|
|
SHA_CTX c;
|
|
|
|
|
|
|
|
/* Generate the header */
|
|
|
|
*hdrlen = sprintf((char *)hdr, "%s %lu", type, len)+1;
|
|
|
|
|
|
|
|
/* Sha1.. */
|
|
|
|
SHA1_Init(&c);
|
|
|
|
SHA1_Update(&c, hdr, *hdrlen);
|
|
|
|
SHA1_Update(&c, buf, len);
|
|
|
|
SHA1_Final(sha1, &c);
|
|
|
|
}
|
|
|
|
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
/*
|
|
|
|
* Link the tempfile to the final place, possibly creating the
|
|
|
|
* last directory level as you do so.
|
|
|
|
*
|
|
|
|
* Returns the errno on failure, 0 on success.
|
|
|
|
*/
|
2006-09-01 09:17:47 +02:00
|
|
|
static int link_temp_to_file(const char *tmpfile, const char *filename)
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
{
|
|
|
|
int ret;
|
2006-07-12 07:00:16 +02:00
|
|
|
char *dir;
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
|
|
|
|
if (!link(tmpfile, filename))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/*
|
2006-07-12 07:00:16 +02:00
|
|
|
* Try to mkdir the last path component if that failed.
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
*
|
|
|
|
* Re-try the "link()" regardless of whether the mkdir
|
|
|
|
* succeeds, since a race might mean that somebody
|
|
|
|
* else succeeded.
|
|
|
|
*/
|
|
|
|
ret = errno;
|
2006-07-12 07:00:16 +02:00
|
|
|
dir = strrchr(filename, '/');
|
|
|
|
if (dir) {
|
|
|
|
*dir = 0;
|
|
|
|
mkdir(filename, 0777);
|
|
|
|
if (adjust_shared_perm(filename))
|
|
|
|
return -2;
|
|
|
|
*dir = '/';
|
|
|
|
if (!link(tmpfile, filename))
|
|
|
|
return 0;
|
|
|
|
ret = errno;
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Move the just written object into its final resting place
|
|
|
|
*/
|
2006-09-01 09:17:47 +02:00
|
|
|
int move_temp_to_file(const char *tmpfile, const char *filename)
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
{
|
|
|
|
int ret = link_temp_to_file(tmpfile, filename);
|
2005-10-26 19:27:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Coda hack - coda doesn't like cross-directory links,
|
|
|
|
* so we fall back to a rename, which will mean that it
|
|
|
|
* won't be able to check collisions, but that's not a
|
|
|
|
* big deal.
|
|
|
|
*
|
|
|
|
* The same holds for FAT formatted media.
|
|
|
|
*
|
|
|
|
* When this succeeds, we just return 0. We have nothing
|
|
|
|
* left to unlink.
|
|
|
|
*/
|
|
|
|
if (ret && ret != EEXIST) {
|
|
|
|
if (!rename(tmpfile, filename))
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
return 0;
|
2005-10-26 01:41:20 +02:00
|
|
|
ret = errno;
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
}
|
|
|
|
unlink(tmpfile);
|
|
|
|
if (ret) {
|
|
|
|
if (ret != EEXIST) {
|
2005-12-15 08:47:30 +01:00
|
|
|
fprintf(stderr, "unable to write sha1 filename %s: %s\n", filename, strerror(ret));
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
/* FIXME!!! Collision check here ? */
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
static int write_buffer(int fd, const void *buf, size_t len)
|
|
|
|
{
|
|
|
|
while (len) {
|
|
|
|
ssize_t size;
|
|
|
|
|
|
|
|
size = write(fd, buf, len);
|
|
|
|
if (!size)
|
|
|
|
return error("file write: disk full");
|
|
|
|
if (size < 0) {
|
|
|
|
if (errno == EINTR || errno == EAGAIN)
|
|
|
|
continue;
|
|
|
|
return error("file write error (%s)", strerror(errno));
|
|
|
|
}
|
|
|
|
len -= size;
|
2006-06-18 17:18:09 +02:00
|
|
|
buf = (char *) buf + size;
|
2006-05-24 17:30:54 +02:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-07-11 21:48:08 +02:00
|
|
|
static int write_binary_header(unsigned char *hdr, enum object_type type, unsigned long len)
|
|
|
|
{
|
|
|
|
int hdr_len;
|
|
|
|
unsigned char c;
|
|
|
|
|
|
|
|
c = (type << 4) | (len & 15);
|
|
|
|
len >>= 4;
|
|
|
|
hdr_len = 1;
|
|
|
|
while (len) {
|
|
|
|
*hdr++ = c | 0x80;
|
|
|
|
hdr_len++;
|
|
|
|
c = (len & 0x7f);
|
|
|
|
len >>= 7;
|
|
|
|
}
|
|
|
|
*hdr = c;
|
|
|
|
return hdr_len;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void setup_object_header(z_stream *stream, const char *type, unsigned long len)
|
|
|
|
{
|
|
|
|
int obj_type, hdr;
|
|
|
|
|
|
|
|
if (use_legacy_headers) {
|
|
|
|
while (deflate(stream, 0) == Z_OK)
|
|
|
|
/* nothing */;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
if (!strcmp(type, blob_type))
|
|
|
|
obj_type = OBJ_BLOB;
|
|
|
|
else if (!strcmp(type, tree_type))
|
|
|
|
obj_type = OBJ_TREE;
|
|
|
|
else if (!strcmp(type, commit_type))
|
|
|
|
obj_type = OBJ_COMMIT;
|
|
|
|
else if (!strcmp(type, tag_type))
|
|
|
|
obj_type = OBJ_TAG;
|
|
|
|
else
|
|
|
|
die("trying to generate bogus object of type '%s'", type);
|
|
|
|
hdr = write_binary_header(stream->next_out, obj_type, len);
|
|
|
|
stream->total_out = hdr;
|
|
|
|
stream->next_out += hdr;
|
|
|
|
stream->avail_out -= hdr;
|
|
|
|
}
|
|
|
|
|
2006-10-14 12:45:36 +02:00
|
|
|
int hash_sha1_file(void *buf, unsigned long len, const char *type,
|
|
|
|
unsigned char *sha1)
|
|
|
|
{
|
|
|
|
unsigned char hdr[50];
|
|
|
|
int hdrlen;
|
|
|
|
write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-05-18 14:14:09 +02:00
|
|
|
int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
|
2005-04-18 22:04:43 +02:00
|
|
|
{
|
|
|
|
int size;
|
2005-05-18 14:14:09 +02:00
|
|
|
unsigned char *compressed;
|
2005-04-18 22:04:43 +02:00
|
|
|
z_stream stream;
|
|
|
|
unsigned char sha1[20];
|
2005-04-20 18:28:05 +02:00
|
|
|
char *filename;
|
2005-05-03 20:46:16 +02:00
|
|
|
static char tmpfile[PATH_MAX];
|
2005-05-18 14:14:09 +02:00
|
|
|
unsigned char hdr[50];
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
int fd, hdrlen;
|
2005-04-25 19:19:53 +02:00
|
|
|
|
2005-06-28 04:03:13 +02:00
|
|
|
/* Normally if we have it in the pack then we do not bother writing
|
|
|
|
* it out into .git/objects/??/?{38} file.
|
|
|
|
*/
|
2006-10-15 14:02:03 +02:00
|
|
|
write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
|
|
|
|
filename = sha1_file_name(sha1);
|
2005-04-20 18:28:05 +02:00
|
|
|
if (returnsha1)
|
2006-08-23 08:49:00 +02:00
|
|
|
hashcpy(returnsha1, sha1);
|
2005-06-28 04:03:13 +02:00
|
|
|
if (has_sha1_file(sha1))
|
|
|
|
return 0;
|
2005-05-03 20:46:16 +02:00
|
|
|
fd = open(filename, O_RDONLY);
|
|
|
|
if (fd >= 0) {
|
2005-04-20 18:28:05 +02:00
|
|
|
/*
|
2005-05-03 20:46:16 +02:00
|
|
|
* FIXME!!! We might do collision checking here, but we'd
|
|
|
|
* need to uncompress the old file and check it. Later.
|
2005-04-20 18:28:05 +02:00
|
|
|
*/
|
2005-05-03 20:46:16 +02:00
|
|
|
close(fd);
|
2005-04-20 18:28:05 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-05-03 20:46:16 +02:00
|
|
|
if (errno != ENOENT) {
|
2005-12-15 08:47:30 +01:00
|
|
|
fprintf(stderr, "sha1 file %s: %s\n", filename, strerror(errno));
|
2005-05-03 20:46:16 +02:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
|
2005-05-07 09:38:04 +02:00
|
|
|
|
2005-05-03 20:46:16 +02:00
|
|
|
fd = mkstemp(tmpfile);
|
|
|
|
if (fd < 0) {
|
2005-12-15 08:47:30 +01:00
|
|
|
fprintf(stderr, "unable to create temporary sha1 filename %s: %s\n", tmpfile, strerror(errno));
|
2005-05-03 20:46:16 +02:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2005-04-18 22:04:43 +02:00
|
|
|
/* Set it up */
|
|
|
|
memset(&stream, 0, sizeof(stream));
|
2006-07-03 22:11:47 +02:00
|
|
|
deflateInit(&stream, zlib_compression_level);
|
2006-07-11 21:48:08 +02:00
|
|
|
size = 8 + deflateBound(&stream, len+hdrlen);
|
2005-04-26 21:00:58 +02:00
|
|
|
compressed = xmalloc(size);
|
2005-04-18 22:04:43 +02:00
|
|
|
|
|
|
|
/* Compress it */
|
|
|
|
stream.next_out = compressed;
|
|
|
|
stream.avail_out = size;
|
2005-04-25 19:19:53 +02:00
|
|
|
|
|
|
|
/* First header.. */
|
|
|
|
stream.next_in = hdr;
|
|
|
|
stream.avail_in = hdrlen;
|
2006-07-11 21:48:08 +02:00
|
|
|
setup_object_header(&stream, type, len);
|
2005-04-25 19:19:53 +02:00
|
|
|
|
|
|
|
/* Then the data itself.. */
|
|
|
|
stream.next_in = buf;
|
|
|
|
stream.avail_in = len;
|
2005-04-18 22:04:43 +02:00
|
|
|
while (deflate(&stream, Z_FINISH) == Z_OK)
|
|
|
|
/* nothing */;
|
|
|
|
deflateEnd(&stream);
|
|
|
|
size = stream.total_out;
|
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
if (write_buffer(fd, compressed, size) < 0)
|
|
|
|
die("unable to write sha1 file");
|
2005-05-03 20:46:16 +02:00
|
|
|
fchmod(fd, 0444);
|
2005-04-20 18:28:05 +02:00
|
|
|
close(fd);
|
2005-05-07 10:27:00 +02:00
|
|
|
free(compressed);
|
2005-04-18 22:04:43 +02:00
|
|
|
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
return move_temp_to_file(tmpfile, filename);
|
2005-04-18 22:04:43 +02:00
|
|
|
}
|
2005-04-24 03:47:23 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
/*
|
|
|
|
* We need to unpack and recompress the object for writing
|
|
|
|
* it out to a different file.
|
|
|
|
*/
|
|
|
|
static void *repack_object(const unsigned char *sha1, unsigned long *objsize)
|
2005-07-11 00:25:38 +02:00
|
|
|
{
|
2006-05-24 17:30:54 +02:00
|
|
|
size_t size;
|
2005-07-11 00:25:38 +02:00
|
|
|
z_stream stream;
|
2006-05-24 17:30:54 +02:00
|
|
|
unsigned char *unpacked;
|
|
|
|
unsigned long len;
|
|
|
|
char type[20];
|
|
|
|
char hdr[50];
|
|
|
|
int hdrlen;
|
|
|
|
void *buf;
|
2005-08-08 20:45:36 +02:00
|
|
|
|
2006-07-10 08:57:51 +02:00
|
|
|
/* need to unpack and recompress it by itself */
|
2006-05-24 17:30:54 +02:00
|
|
|
unpacked = read_packed_sha1(sha1, type, &len);
|
2005-07-11 00:25:38 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
hdrlen = sprintf(hdr, "%s %lu", type, len) + 1;
|
2005-07-11 00:25:38 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
/* Set it up */
|
|
|
|
memset(&stream, 0, sizeof(stream));
|
2006-07-03 22:11:47 +02:00
|
|
|
deflateInit(&stream, zlib_compression_level);
|
2006-05-24 17:30:54 +02:00
|
|
|
size = deflateBound(&stream, len + hdrlen);
|
|
|
|
buf = xmalloc(size);
|
2005-07-11 00:25:38 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
/* Compress it */
|
|
|
|
stream.next_out = buf;
|
|
|
|
stream.avail_out = size;
|
2005-07-11 00:25:38 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
/* First header.. */
|
|
|
|
stream.next_in = (void *)hdr;
|
|
|
|
stream.avail_in = hdrlen;
|
|
|
|
while (deflate(&stream, 0) == Z_OK)
|
|
|
|
/* nothing */;
|
2005-08-08 20:45:36 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
/* Then the data itself.. */
|
|
|
|
stream.next_in = unpacked;
|
|
|
|
stream.avail_in = len;
|
|
|
|
while (deflate(&stream, Z_FINISH) == Z_OK)
|
|
|
|
/* nothing */;
|
|
|
|
deflateEnd(&stream);
|
|
|
|
free(unpacked);
|
2005-08-08 20:45:36 +02:00
|
|
|
|
2006-05-24 17:30:54 +02:00
|
|
|
*objsize = stream.total_out;
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
int write_sha1_to_fd(int fd, const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int retval;
|
|
|
|
unsigned long objsize;
|
2006-07-18 00:04:47 +02:00
|
|
|
void *buf = map_sha1_file(sha1, &objsize);
|
2006-05-24 17:30:54 +02:00
|
|
|
|
|
|
|
if (buf) {
|
|
|
|
retval = write_buffer(fd, buf, objsize);
|
|
|
|
munmap(buf, objsize);
|
|
|
|
return retval;
|
|
|
|
}
|
|
|
|
|
|
|
|
buf = repack_object(sha1, &objsize);
|
|
|
|
retval = write_buffer(fd, buf, objsize);
|
|
|
|
free(buf);
|
|
|
|
return retval;
|
2005-07-11 00:25:38 +02:00
|
|
|
}
|
|
|
|
|
2005-08-03 01:46:29 +02:00
|
|
|
int write_sha1_from_fd(const unsigned char *sha1, int fd, char *buffer,
|
|
|
|
size_t bufsize, size_t *bufposn)
|
2005-04-24 03:47:23 +02:00
|
|
|
{
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
char tmpfile[PATH_MAX];
|
2005-04-24 03:47:23 +02:00
|
|
|
int local;
|
|
|
|
z_stream stream;
|
|
|
|
unsigned char real_sha1[20];
|
2005-05-18 14:14:09 +02:00
|
|
|
unsigned char discard[4096];
|
2005-04-24 03:47:23 +02:00
|
|
|
int ret;
|
|
|
|
SHA_CTX c;
|
|
|
|
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
|
2005-04-24 03:47:23 +02:00
|
|
|
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
local = mkstemp(tmpfile);
|
2005-04-24 03:47:23 +02:00
|
|
|
if (local < 0)
|
2006-02-23 02:47:10 +01:00
|
|
|
return error("Couldn't open %s for %s",
|
|
|
|
tmpfile, sha1_to_hex(sha1));
|
2005-04-24 03:47:23 +02:00
|
|
|
|
|
|
|
memset(&stream, 0, sizeof(stream));
|
|
|
|
|
|
|
|
inflateInit(&stream);
|
|
|
|
|
|
|
|
SHA1_Init(&c);
|
|
|
|
|
|
|
|
do {
|
|
|
|
ssize_t size;
|
2005-08-03 01:46:29 +02:00
|
|
|
if (*bufposn) {
|
|
|
|
stream.avail_in = *bufposn;
|
2005-08-09 22:54:40 +02:00
|
|
|
stream.next_in = (unsigned char *) buffer;
|
2005-08-03 01:46:29 +02:00
|
|
|
do {
|
|
|
|
stream.next_out = discard;
|
|
|
|
stream.avail_out = sizeof(discard);
|
|
|
|
ret = inflate(&stream, Z_SYNC_FLUSH);
|
|
|
|
SHA1_Update(&c, discard, sizeof(discard) -
|
|
|
|
stream.avail_out);
|
|
|
|
} while (stream.avail_in && ret == Z_OK);
|
2006-05-24 17:30:54 +02:00
|
|
|
if (write_buffer(local, buffer, *bufposn - stream.avail_in) < 0)
|
|
|
|
die("unable to write sha1 file");
|
2005-08-03 01:46:29 +02:00
|
|
|
memmove(buffer, buffer + *bufposn - stream.avail_in,
|
|
|
|
stream.avail_in);
|
|
|
|
*bufposn = stream.avail_in;
|
|
|
|
if (ret != Z_OK)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
size = read(fd, buffer + *bufposn, bufsize - *bufposn);
|
2005-04-24 03:47:23 +02:00
|
|
|
if (size <= 0) {
|
|
|
|
close(local);
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
unlink(tmpfile);
|
2005-04-24 03:47:23 +02:00
|
|
|
if (!size)
|
|
|
|
return error("Connection closed?");
|
|
|
|
perror("Reading from connection");
|
|
|
|
return -1;
|
|
|
|
}
|
2005-08-03 01:46:29 +02:00
|
|
|
*bufposn += size;
|
|
|
|
} while (1);
|
2005-04-24 03:47:23 +02:00
|
|
|
inflateEnd(&stream);
|
|
|
|
|
|
|
|
close(local);
|
|
|
|
SHA1_Final(real_sha1, &c);
|
|
|
|
if (ret != Z_STREAM_END) {
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
unlink(tmpfile);
|
2005-04-24 03:47:23 +02:00
|
|
|
return error("File %s corrupted", sha1_to_hex(sha1));
|
|
|
|
}
|
2006-08-17 20:54:57 +02:00
|
|
|
if (hashcmp(sha1, real_sha1)) {
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
unlink(tmpfile);
|
2006-02-23 02:47:10 +01:00
|
|
|
return error("File %s has bad hash", sha1_to_hex(sha1));
|
2005-04-24 03:47:23 +02:00
|
|
|
}
|
Create object subdirectories on demand
This makes it possible to have a "sparse" git object subdirectory
structure, something that has become much more attractive now that people
use pack-files all the time.
As a result of pack-files, a git object directory doesn't necessarily have
any individual objects lying around, and in that case it's just wasting
space to keep the empty first-level object directories around: on many
filesystems the 256 empty directories will be aboue 1MB of diskspace.
Even more importantly, after you re-pack a project that _used_ to be
unpacked, you could be left with huge directories that no longer contain
anything, but that waste space and take time to look through.
With this change, "git prune-packed" can just do an rmdir() on the
directories, and they'll get removed if empty, and re-created on demand.
This patch also tries to fix up "write_sha1_from_fd()" to use the new
common infrastructure for creating the object files, closing a hole where
we might otherwise leave half-written objects in the object database.
[jc: I unoptimized the part that really removes the fan-out directories
to ease transition. init-db still wastes 1MB of diskspace to hold 256
empty fan-outs, and prune-packed rmdir()'s the grown but empty directories,
but runs mkdir() immediately after that -- reducing the saving from 150KB
to 146KB. These parts will be re-introduced when everybody has the
on-demand capability.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-10-09 00:54:01 +02:00
|
|
|
|
|
|
|
return move_temp_to_file(tmpfile, sha1_file_name(sha1));
|
2005-04-24 03:47:23 +02:00
|
|
|
}
|
|
|
|
|
2005-08-01 02:53:44 +02:00
|
|
|
int has_pack_index(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
struct stat st;
|
|
|
|
if (stat(sha1_pack_index_name(sha1), &st))
|
|
|
|
return 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
int has_pack_file(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
struct stat st;
|
|
|
|
if (stat(sha1_pack_name(sha1), &st))
|
|
|
|
return 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2006-09-06 11:12:09 +02:00
|
|
|
int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed)
|
2005-07-03 22:06:36 +02:00
|
|
|
{
|
|
|
|
struct pack_entry e;
|
2006-09-06 11:12:09 +02:00
|
|
|
return find_pack_entry(sha1, &e, ignore_packed);
|
2005-07-03 22:06:36 +02:00
|
|
|
}
|
|
|
|
|
2005-04-24 03:47:23 +02:00
|
|
|
int has_sha1_file(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
struct stat st;
|
2005-06-27 12:35:33 +02:00
|
|
|
struct pack_entry e;
|
|
|
|
|
2006-09-06 11:12:09 +02:00
|
|
|
if (find_pack_entry(sha1, &e, NULL))
|
2005-06-27 12:35:33 +02:00
|
|
|
return 1;
|
2005-07-11 09:00:55 +02:00
|
|
|
return find_sha1_file(sha1, &st) ? 1 : 0;
|
2005-04-24 03:47:23 +02:00
|
|
|
}
|
2005-05-02 08:45:49 +02:00
|
|
|
|
2006-05-23 20:19:04 +02:00
|
|
|
/*
|
|
|
|
* reads from fd as long as possible into a supplied buffer of size bytes.
|
2006-07-10 07:50:18 +02:00
|
|
|
* If necessary the buffer's size is increased using realloc()
|
2006-05-23 20:19:04 +02:00
|
|
|
*
|
|
|
|
* returns 0 if anything went fine and -1 otherwise
|
|
|
|
*
|
|
|
|
* NOTE: both buf and size may change, but even when -1 is returned
|
|
|
|
* you still have to free() it yourself.
|
|
|
|
*/
|
|
|
|
int read_pipe(int fd, char** return_buf, unsigned long* return_size)
|
2005-12-10 23:25:24 +01:00
|
|
|
{
|
2006-05-23 20:19:04 +02:00
|
|
|
char* buf = *return_buf;
|
|
|
|
unsigned long size = *return_size;
|
|
|
|
int iret;
|
2005-12-10 23:25:24 +01:00
|
|
|
unsigned long off = 0;
|
2006-05-23 20:19:04 +02:00
|
|
|
|
2005-12-10 23:25:24 +01:00
|
|
|
do {
|
2006-05-23 20:19:04 +02:00
|
|
|
iret = xread(fd, buf + off, size - off);
|
2005-12-10 23:25:24 +01:00
|
|
|
if (iret > 0) {
|
|
|
|
off += iret;
|
|
|
|
if (off == size) {
|
|
|
|
size *= 2;
|
2006-08-26 16:16:18 +02:00
|
|
|
buf = xrealloc(buf, size);
|
2005-12-10 23:25:24 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
} while (iret > 0);
|
2006-05-23 20:19:04 +02:00
|
|
|
|
|
|
|
*return_buf = buf;
|
|
|
|
*return_size = off;
|
|
|
|
|
|
|
|
if (iret < 0)
|
|
|
|
return -1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int index_pipe(unsigned char *sha1, int fd, const char *type, int write_object)
|
|
|
|
{
|
|
|
|
unsigned long size = 4096;
|
2006-09-01 00:32:39 +02:00
|
|
|
char *buf = xmalloc(size);
|
2006-05-23 20:19:04 +02:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (read_pipe(fd, &buf, &size)) {
|
2005-12-10 23:25:24 +01:00
|
|
|
free(buf);
|
|
|
|
return -1;
|
|
|
|
}
|
2006-05-23 20:19:04 +02:00
|
|
|
|
2005-12-10 23:25:24 +01:00
|
|
|
if (!type)
|
2006-04-02 14:44:09 +02:00
|
|
|
type = blob_type;
|
2005-12-10 23:25:24 +01:00
|
|
|
if (write_object)
|
2006-05-23 20:19:04 +02:00
|
|
|
ret = write_sha1_file(buf, size, type, sha1);
|
2006-10-14 12:45:36 +02:00
|
|
|
else
|
|
|
|
ret = hash_sha1_file(buf, size, type, sha1);
|
2005-12-10 23:25:24 +01:00
|
|
|
free(buf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2005-07-09 01:51:55 +02:00
|
|
|
int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object, const char *type)
|
2005-05-02 08:45:49 +02:00
|
|
|
{
|
|
|
|
unsigned long size = st->st_size;
|
2005-05-03 20:46:16 +02:00
|
|
|
void *buf;
|
|
|
|
int ret;
|
2005-05-02 08:45:49 +02:00
|
|
|
|
2005-05-03 20:46:16 +02:00
|
|
|
buf = "";
|
2005-05-02 08:45:49 +02:00
|
|
|
if (size)
|
2005-05-03 20:46:16 +02:00
|
|
|
buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
|
2005-05-02 08:45:49 +02:00
|
|
|
close(fd);
|
2005-07-29 16:49:14 +02:00
|
|
|
if (buf == MAP_FAILED)
|
2005-05-02 08:45:49 +02:00
|
|
|
return -1;
|
|
|
|
|
2005-07-09 01:51:55 +02:00
|
|
|
if (!type)
|
2006-04-02 14:44:09 +02:00
|
|
|
type = blob_type;
|
2005-07-09 01:51:55 +02:00
|
|
|
if (write_object)
|
|
|
|
ret = write_sha1_file(buf, size, type, sha1);
|
2006-10-14 12:45:36 +02:00
|
|
|
else
|
|
|
|
ret = hash_sha1_file(buf, size, type, sha1);
|
2005-05-03 20:46:16 +02:00
|
|
|
if (size)
|
|
|
|
munmap(buf, size);
|
|
|
|
return ret;
|
2005-05-02 08:45:49 +02:00
|
|
|
}
|
2005-10-07 12:42:00 +02:00
|
|
|
|
|
|
|
int index_path(unsigned char *sha1, const char *path, struct stat *st, int write_object)
|
|
|
|
{
|
|
|
|
int fd;
|
|
|
|
char *target;
|
|
|
|
|
|
|
|
switch (st->st_mode & S_IFMT) {
|
|
|
|
case S_IFREG:
|
|
|
|
fd = open(path, O_RDONLY);
|
|
|
|
if (fd < 0)
|
|
|
|
return error("open(\"%s\"): %s", path,
|
|
|
|
strerror(errno));
|
|
|
|
if (index_fd(sha1, fd, st, write_object, NULL) < 0)
|
|
|
|
return error("%s: failed to insert into database",
|
|
|
|
path);
|
|
|
|
break;
|
|
|
|
case S_IFLNK:
|
|
|
|
target = xmalloc(st->st_size+1);
|
|
|
|
if (readlink(path, target, st->st_size+1) != st->st_size) {
|
|
|
|
char *errstr = strerror(errno);
|
|
|
|
free(target);
|
|
|
|
return error("readlink(\"%s\"): %s", path,
|
|
|
|
errstr);
|
|
|
|
}
|
2006-10-14 12:45:36 +02:00
|
|
|
if (!write_object)
|
|
|
|
hash_sha1_file(target, st->st_size, blob_type, sha1);
|
|
|
|
else if (write_sha1_file(target, st->st_size, blob_type, sha1))
|
2005-10-07 12:42:00 +02:00
|
|
|
return error("%s: failed to insert into database",
|
|
|
|
path);
|
|
|
|
free(target);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return error("%s: unsupported file type", path);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|