Commit e83c5163 authored by Linus Torvalds's avatar Linus Torvalds

Initial revision of "git", the information manager from hell

PROG=update-cache show-diff init-db write-tree read-tree commit-tree cat-file
all: $(PROG)
install: $(PROG)
install $(PROG) $(HOME)/bin/
LIBS= -lssl
init-db: init-db.o
update-cache: update-cache.o read-cache.o
$(CC) $(CFLAGS) -o update-cache update-cache.o read-cache.o $(LIBS)
show-diff: show-diff.o read-cache.o
$(CC) $(CFLAGS) -o show-diff show-diff.o read-cache.o $(LIBS)
write-tree: write-tree.o read-cache.o
$(CC) $(CFLAGS) -o write-tree write-tree.o read-cache.o $(LIBS)
read-tree: read-tree.o read-cache.o
$(CC) $(CFLAGS) -o read-tree read-tree.o read-cache.o $(LIBS)
commit-tree: commit-tree.o read-cache.o
$(CC) $(CFLAGS) -o commit-tree commit-tree.o read-cache.o $(LIBS)
cat-file: cat-file.o read-cache.o
$(CC) $(CFLAGS) -o cat-file cat-file.o read-cache.o $(LIBS)
read-cache.o: cache.h
show-diff.o: cache.h
rm -f *.o $(PROG) temp_git_file_*
backup: clean
cd .. ; tar czvf dircache.tar.gz dir-cache
GIT - the stupid content tracker
"git" can mean anything, depending on your mood.
- random three-letter combination that is pronounceable, and not
actually used by any common UNIX command. The fact that it is a
mispronounciation of "get" may or may not be relevant.
- stupid. contemptible and despicable. simple. Take your pick from the
dictionary of slang.
- "global information tracker": you're in a good mood, and it actually
works for you. Angels sing, and a light suddenly fills the room.
- "goddamn idiotic truckload of sh*t": when it breaks
This is a stupid (but extremely fast) directory content manager. It
doesn't do a whole lot, but what it _does_ do is track directory
contents efficiently.
There are two object abstractions: the "object database", and the
"current directory cache".
The Object Database (SHA1_FILE_DIRECTORY)
The object database is literally just a content-addressable collection
of objects. All objects are named by their content, which is
approximated by the SHA1 hash of the object itself. Objects may refer
to other objects (by referencing their SHA1 hash), and so you can build
up a hierarchy of objects.
There are several kinds of objects in the content-addressable collection
database. They are all in deflated with zlib, and start off with a tag
of their type, and size information about the data. The SHA1 hash is
always the hash of the _compressed_ object, not the original one.
In particular, the consistency of an object can always be tested
independently of the contents or the type of the object: all objects can
be validated by verifying that (a) their hashes match the content of the
file and (b) the object successfully inflates to a stream of bytes that
forms a sequence of <ascii tag without space> + <space> + <ascii decimal
size> + <byte\0> + <binary object data>.
BLOB: A "blob" object is nothing but a binary blob of data, and doesn't
refer to anything else. There is no signature or any other verification
of the data, so while the object is consistent (it _is_ indexed by its
sha1 hash, so the data itself is certainly correct), it has absolutely
no other attributes. No name associations, no permissions. It is
purely a blob of data (ie normally "file contents").
TREE: The next hierarchical object type is the "tree" object. A tree
object is a list of permission/name/blob data, sorted by name. In other
words the tree object is uniquely determined by the set contents, and so
two separate but identical trees will always share the exact same
Again, a "tree" object is just a pure data abstraction: it has no
history, no signatures, no verification of validity, except that the
contents are again protected by the hash itself. So you can trust the
contents of a tree, the same way you can trust the contents of a blob,
but you don't know where those contents _came_ from.
Side note on trees: since a "tree" object is a sorted list of
"filename+content", you can create a diff between two trees without
actually having to unpack two trees. Just ignore all common parts, and
your diff will look right. In other words, you can effectively (and
efficiently) tell the difference between any two random trees by O(n)
where "n" is the size of the difference, rather than the size of the
Side note 2 on trees: since the name of a "blob" depends entirely and
exclusively on its contents (ie there are no names or permissions
involved), you can see trivial renames or permission changes by noticing
that the blob stayed the same. However, renames with data changes need
a smarter "diff" implementation.
CHANGESET: The "changeset" object is an object that introduces the
notion of history into the picture. In contrast to the other objects,
it doesn't just describe the physical state of a tree, it describes how
we got there, and why.
A "changeset" is defined by the tree-object that it results in, the
parent changesets (zero, one or more) that led up to that point, and a
comment on what happened. Again, a changeset is not trusted per se:
the contents are well-defined and "safe" due to the cryptographically
strong signatures at all levels, but there is no reason to believe that
the tree is "good" or that the merge information makes sense. The
parents do not have to actually have any relationship with the result,
for example.
Note on changesets: unlike real SCM's, changesets do not contain rename
information or file mode chane information. All of that is implicit in
the trees involved (the result tree, and the result trees of the
parents), and describing that makes no sense in this idiotic file
TRUST: The notion of "trust" is really outside the scope of "git", but
it's worth noting a few things. First off, since everything is hashed
with SHA1, you _can_ trust that an object is intact and has not been
messed with by external sources. So the name of an object uniquely
identifies a known state - just not a state that you may want to trust.
Furthermore, since the SHA1 signature of a changeset refers to the
SHA1 signatures of the tree it is associated with and the signatures
of the parent, a single named changeset specifies uniquely a whole
set of history, with full contents. You can't later fake any step of
the way once you have the name of a changeset.
So to introduce some real trust in the system, the only thing you need
to do is to digitally sign just _one_ special note, which includes the
name of a top-level changeset. Your digital signature shows others that
you trust that changeset, and the immutability of the history of
changesets tells others that they can trust the whole history.
In other words, you can easily validate a whole archive by just sending
out a single email that tells the people the name (SHA1 hash) of the top
changeset, and digitally sign that email using something like GPG/PGP.
In particular, you can also have a separate archive of "trust points" or
tags, which document your (and other peoples) trust. You may, of
course, archive these "certificates of trust" using "git" itself, but
it's not something "git" does for you.
Another way of saying the same thing: "git" itself only handles content
integrity, the trust has to come from outside.
Current Directory Cache (".dircache/index")
The "current directory cache" is a simple binary file, which contains an
efficient representation of a virtual directory content at some random
time. It does so by a simple array that associates a set of names,
dates, permissions and content (aka "blob") objects together. The cache
is always kept ordered by name, and names are unique at any point in
time, but the cache has no long-term meaning, and can be partially
updated at any time.
In particular, the "current directory cache" certainly does not need to
be consistent with the current directory contents, but it has two very
important attributes:
(a) it can re-generate the full state it caches (not just the directory
structure: through the "blob" object it can regenerate the data too)
As a special case, there is a clear and unambiguous one-way mapping
from a current directory cache to a "tree object", which can be
efficiently created from just the current directory cache without
actually looking at any other data. So a directory cache at any
one time uniquely specifies one and only one "tree" object (but
has additional data to make it easy to match up that tree object
with what has happened in the directory)
(b) it has efficient methods for finding inconsistencies between that
cached state ("tree object waiting to be instantiated") and the
current state.
Those are the two ONLY things that the directory cache does. It's a
cache, and the normal operation is to re-generate it completely from a
known tree object, or update/compare it with a live tree that is being
developed. If you blow the directory cache away entirely, you haven't
lost any information as long as you have the name of the tree that it
(But directory caches can also have real information in them: in
particular, they can have the representation of an intermediate tree
that has not yet been instantiated. So they do have meaning and usage
outside of caching - in one sense you can think of the current directory
cache as being the "work in progress" towards a tree commit).
#ifndef CACHE_H
#define CACHE_H
#include <stdio.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <stddef.h>
#include <stdlib.h>
#include <stdarg.h>
#include <errno.h>
#include <sys/mman.h>
#include <openssl/sha.h>
#include <zlib.h>
* Basic data structures for the directory cache
* NOTE NOTE NOTE! This is all in the native CPU byte format. It's
* not even trying to be portable. It's trying to be efficient. It's
* just a cache, after all.
#define CACHE_SIGNATURE 0x44495243 /* "DIRC" */
struct cache_header {
unsigned int signature;
unsigned int version;
unsigned int entries;
unsigned char sha1[20];
* The "cache_time" is just the low 32 bits of the
* time. It doesn't matter if it overflows - we only
* check it for equality in the 32 bits we save.
struct cache_time {
unsigned int sec;
unsigned int nsec;
* dev/ino/uid/gid/size are also just tracked to the low 32 bits
* Again - this is just a (very strong in practice) heuristic that
* the inode hasn't changed.
struct cache_entry {
struct cache_time ctime;
struct cache_time mtime;
unsigned int st_dev;
unsigned int st_ino;
unsigned int st_mode;
unsigned int st_uid;
unsigned int st_gid;
unsigned int st_size;
unsigned char sha1[20];
unsigned short namelen;
unsigned char name[0];
const char *sha1_file_directory;
struct cache_entry **active_cache;
unsigned int active_nr, active_alloc;
#define DEFAULT_DB_ENVIRONMENT ".dircache/objects"
#define cache_entry_size(len) ((offsetof(struct cache_entry,name) + (len) + 8) & ~7)
#define ce_size(ce) cache_entry_size((ce)->namelen)
#define alloc_nr(x) (((x)+16)*3/2)
/* Initialize the cache information */
extern int read_cache(void);
/* Return a statically allocated filename matching the sha1 signature */
extern char *sha1_file_name(unsigned char *sha1);
/* Write a memory buffer out to the sha file */
extern int write_sha1_buffer(unsigned char *sha1, void *buf, unsigned int size);
/* Read and unpack a sha1 file into memory, write memory to a sha1 file */
extern void * read_sha1_file(unsigned char *sha1, char *type, unsigned long *size);
extern int write_sha1_file(char *buf, unsigned len);
/* Convert to/from hex/sha1 representation */
extern int get_sha1_hex(char *hex, unsigned char *sha1);
extern char *sha1_to_hex(unsigned char *sha1); /* static buffer! */
/* General helper functions */
extern void usage(const char *err);
#endif /* CACHE_H */
#include "cache.h"
int main(int argc, char **argv)
unsigned char sha1[20];
char type[20];
void *buf;
unsigned long size;
char template[] = "temp_git_file_XXXXXX";
int fd;
if (argc != 2 || get_sha1_hex(argv[1], sha1))
usage("cat-file: cat-file <sha1>");
buf = read_sha1_file(sha1, type, &size);
if (!buf)
fd = mkstemp(template);
if (fd < 0)
usage("unable to create tempfile");
if (write(fd, buf, size) != size)
strcpy(type, "bad");
printf("%s: %s\n", template, type);
#include "cache.h"
#include <pwd.h>
#include <time.h>
#define BLOCKING (1ul << 14)
#define ORIG_OFFSET (40)
* Leave space at the beginning to insert the tag
* once we know how big things are.
* FIXME! Share the code with "write-tree.c"
static void init_buffer(char **bufp, unsigned int *sizep)
char *buf = malloc(BLOCKING);
memset(buf, 0, ORIG_OFFSET);
*sizep = ORIG_OFFSET;
*bufp = buf;
static void add_buffer(char **bufp, unsigned int *sizep, const char *fmt, ...)
char one_line[2048];
va_list args;
int len;
unsigned long alloc, size, newsize;
char *buf;
va_start(args, fmt);
len = vsnprintf(one_line, sizeof(one_line), fmt, args);
size = *sizep;
newsize = size + len;
alloc = (size + 32767) & ~32767;
buf = *bufp;
if (newsize > alloc) {
alloc = (newsize + 32767) & ~32767;
buf = realloc(buf, alloc);
*bufp = buf;
*sizep = newsize;
memcpy(buf + size, one_line, len);
static int prepend_integer(char *buffer, unsigned val, int i)
buffer[--i] = '\0';
do {
buffer[--i] = '0' + (val % 10);
val /= 10;
} while (val);
return i;
static void finish_buffer(char *tag, char **bufp, unsigned int *sizep)
int taglen;
int offset;
char *buf = *bufp;
unsigned int size = *sizep;
offset = prepend_integer(buf, size - ORIG_OFFSET, ORIG_OFFSET);
taglen = strlen(tag);
offset -= taglen;
buf += offset;
size -= offset;
memcpy(buf, tag, taglen);
*bufp = buf;
*sizep = size;
static void remove_special(char *p)
char c;
char *dst = p;
for (;;) {
c = *p;
switch(c) {
case '\n': case '<': case '>':
*dst++ = c;
if (!c)
* Having more than two parents may be strange, but hey, there's
* no conceptual reason why the file format couldn't accept multi-way
* merges. It might be the "union" of several packages, for example.
* I don't really expect that to happen, but this is here to make
* it clear that _conceptually_ it's ok..
#define MAXPARENT (16)
int main(int argc, char **argv)
int i, len;
int parents = 0;
unsigned char tree_sha1[20];
unsigned char parent_sha1[MAXPARENT][20];
char *gecos, *realgecos;
char *email, realemail[1000];
char *date, *realdate;
char comment[1000];
struct passwd *pw;
time_t now;
char *buffer;
unsigned int size;
if (argc < 2 || get_sha1_hex(argv[1], tree_sha1) < 0)
usage("commit-tree <sha1> [-p <sha1>]* < changelog");
for (i = 2; i < argc; i += 2) {
char *a, *b;
a = argv[i]; b = argv[i+1];
if (!b || strcmp(a, "-p") || get_sha1_hex(b, parent_sha1[parents]))
usage("commit-tree <sha1> [-p <sha1>]* < changelog");
if (!parents)
fprintf(stderr, "Committing initial tree %s\n", argv[1]);
pw = getpwuid(getuid());
if (!pw)
usage("You don't exist. Go away!");
realgecos = pw->pw_gecos;
len = strlen(pw->pw_name);
memcpy(realemail, pw->pw_name, len);
realemail[len] = '@';
gethostname(realemail+len+1, sizeof(realemail)-len-1);
realdate = ctime(&now);
gecos = getenv("COMMITTER_NAME") ? : realgecos;
email = getenv("COMMITTER_EMAIL") ? : realemail;
date = getenv("COMMITTER_DATE") ? : realdate;
remove_special(gecos); remove_special(realgecos);
remove_special(email); remove_special(realemail);
remove_special(date); remove_special(realdate);
init_buffer(&buffer, &size);
add_buffer(&buffer, &size, "tree %s\n", sha1_to_hex(tree_sha1));
* NOTE! This ordering means that the same exact tree merged with a
* different order of parents will be a _different_ changeset even
* if everything else stays the same.
for (i = 0; i < parents; i++)
add_buffer(&buffer, &size, "parent %s\n", sha1_to_hex(parent_sha1[i]));
/* Person/date information */
add_buffer(&buffer, &size, "author %s <%s> %s\n", gecos, email, date);
add_buffer(&buffer, &size, "committer %s <%s> %s\n\n", realgecos, realemail, realdate);
/* And add the comment */
while (fgets(comment, sizeof(comment), stdin) != NULL)
add_buffer(&buffer, &size, "%s", comment);
finish_buffer("commit ", &buffer, &size);
write_sha1_file(buffer, size);
return 0;
#include "cache.h"
int main(int argc, char **argv)
char *sha1_dir = getenv(DB_ENVIRONMENT), *path;
int len, i, fd;
if (mkdir(".dircache", 0700) < 0) {
perror("unable to create .dircache");
* If you want to, you can share the DB area with any number of branches.
* That has advantages: you can save space by sharing all the SHA1 objects.
* On the other hand, it might just make lookup slower and messier. You
* be the judge.
sha1_dir = getenv(DB_ENVIRONMENT);
if (sha1_dir) {
struct stat st;
if (!stat(sha1_dir, &st) < 0 && S_ISDIR(st.st_mode))
fprintf(stderr, "DB_ENVIRONMENT set to bad directory %s: ", sha1_dir);
* The default case is to have a DB per managed directory.
fprintf(stderr, "defaulting to private storage area\n");
len = strlen(sha1_dir);
if (mkdir(sha1_dir, 0700) < 0) {
if (errno != EEXIST) {
path = malloc(len + 40);
memcpy(path, sha1_dir, len);
for (i = 0; i < 256; i++) {
sprintf(path+len, "/%02x", i);
if (mkdir(path, 0700) < 0) {
if (errno != EEXIST) {
return 0;
#include "cache.h"
const char *sha1_file_directory = NULL;
struct cache_entry **active_cache = NULL;
unsigned int active_nr = 0, active_alloc = 0;
void usage(const char *err)
fprintf(stderr, "read-tree: %s\n", err);
static unsigned hexval(char c)
if (c >= '0' && c <= '9')
return c - '0';
if (c >= 'a' && c <= 'f')
return c - 'a' + 10;
if (c >= 'A' && c <= 'F')
return c - 'A' + 10;
return ~0;
int get_sha1_hex(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;
char * sha1_to_hex(unsigned char *sha1)
static char buffer[50];
static const char hex[] = "0123456789abcdef";
char *buf = buffer;
int i;
for (i = 0; i < 20; i++) {
unsigned int val = *sha1++;
*buf++ = hex[val >> 4];
*buf++ = hex[val & 0xf];
return buffer;
* NOTE! This returns a statically allocated buffer, so you have to be
* careful about using it. Do a "strdup()" if you need to save the
* filename.
char *sha1_file_name(unsigned char *sha1)
int i;
static char *name, *base;
if (!base) {
char *sha1_file_directory = getenv(DB_ENVIRONMENT) ? : DEFAULT_DB_ENVIRONMENT;
int len = strlen(sha1_file_directory);
base = malloc(len + 60);
memcpy(base, sha1_file_directory, len);
memset(base+len, 0, 60);
base[len] = '/';
base[len+3] = '/';
name = base + len + 1;
for (i = 0; i < 20; i++) {
static char hex[] = "0123456789abcdef";
unsigned int val = sha1[i];
char *pos = name + i*2 + (i > 0);
*pos++ = hex[val >> 4];
*pos = hex[val & 0xf];
return base;
void * read_sha1_file(unsigned char *sha1, char *type, unsigned long *size)
z_stream stream;
char buffer[8192];
struct stat st;
int i, fd, ret, bytes;
void *map, *buf;
char *filename = sha1_file_name(sha1);
fd = open(filename, O_RDONLY);
if (fd < 0) {
return NULL;
if (fstat(fd, &st) < 0) {
return NULL;
map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
if (-1 == (int)(long)map)
return NULL;
/* Get the data stream */
memset(&stream, 0, sizeof(stream));
stream.next_in = map;
stream.avail_in = st.st_size;
stream.next_out = buffer;
stream.avail_out = sizeof(buffer);
ret = inflate(&stream, 0);
if (sscanf(buffer, "%10s %lu", type, size) != 2)
return NULL;
bytes = strlen(buffer) + 1;
buf = malloc(*size);
if (!buf)
return NULL;
memcpy(buf, buffer + bytes, stream.total_out - bytes);
bytes = stream.total_out - bytes;
if (bytes < *size && ret == Z_OK) {
stream.next_out = buf + bytes;
stream.avail_out = *size - bytes;
while (inflate(&stream, Z_FINISH) == Z_OK)
/* nothing */;
return buf;
int write_sha1_file(char *buf, unsigned len)
int size;
char *compressed;
z_stream stream;
unsigned char sha1[20];
/* Set it up */
memset(&stream, 0, sizeof(stream));
deflateInit(&stream, Z_BEST_COMPRESSION);
size = deflateBound(&stream, len);
compressed = malloc(size);
/* Compress it */
stream.next_in = buf;
stream.avail_in = len;
stream.next_out = compressed;
stream.avail_out = size;
while (deflate(&stream, Z_FINISH) == Z_OK)
/* nothing */;
size = stream.total_out;
/* Sha1.. */
SHA1_Update(&c, compressed, size);
SHA1_Final(sha1, &c);
if (write_sha1_buffer(sha1, compressed, size) < 0)
return -1;
printf("%s\n", sha1_to_hex(sha1));
return 0;
int write_sha1_buffer(unsigned char *sha1, void *buf, unsigned int size)
char *filename = sha1_file_name(sha1);
int i, fd;
fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
if (fd < 0)
return (errno == EEXIST) ? 0 : -1;
write(fd, buf, size);
return 0;
static int error(const char * string)
fprintf(stderr, "error: %s\n", string);
return -1;
static int verify_hdr(struct cache_header *hdr, unsigned long size)
unsigned char sha1[20];
if (hdr->signature != CACHE_SIGNATURE)
return error("bad signature");
if (hdr->version != 1)
return error("bad version");
SHA1_Update(&c, hdr, offsetof(struct cache_header, sha1));
SHA1_Update(&c, hdr+1, size - sizeof(*hdr));
SHA1_Final(sha1, &c);
if (memcmp(sha1, hdr->sha1, 20))