[thirdparty/git.git] / sha1_file.c

/*
 * GIT - The information manager from hell
 *
 * Copyright (C) Linus Torvalds, 2005
 *
 * This handles basic git sha1 object files - packing, unpacking,
 * creation etc.
 */
#include <stdarg.h>
#include "cache.h"

const char *sha1_file_directory = NULL;

#ifndef O_NOATIME
#if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
#define O_NOATIME 01000000
#else
#define O_NOATIME 0
#endif
#endif

static unsigned int sha1_file_open_flag = O_NOATIME;

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(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;
}

char * sha1_to_hex(const 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(const 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;
}

int check_sha1_signature(unsigned char *sha1, void *map, unsigned long size, const char *type)
{
	char header[100];
	unsigned char real_sha1[20];
	SHA_CTX c;

	SHA1_Init(&c);
	SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
	SHA1_Update(&c, map, size);
	SHA1_Final(real_sha1, &c);
	return memcmp(sha1, real_sha1, 20) ? -1 : 0;
}

void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
{
	char *filename = sha1_file_name(sha1);
	struct stat st;
	void *map;
	int fd;

	fd = open(filename, O_RDONLY | sha1_file_open_flag);
	if (fd < 0) {
		/* 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:
			perror(filename);
			return NULL;
		}

		/* If it failed once, it will probably fail again. Stop using O_NOATIME */
		sha1_file_open_flag = 0;
	}
	if (fstat(fd, &st) < 0) {
		close(fd);
		return NULL;
	}
	map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
	close(fd);
	if (-1 == (int)(long)map)
		return NULL;
	*size = st.st_size;
	return map;
}

void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
{
	int ret, bytes;
	z_stream stream;
	char buffer[8192];
	char *buf;

	/* Get the data stream */
	memset(&stream, 0, sizeof(stream));
	stream.next_in = map;
	stream.avail_in = mapsize;
	stream.next_out = buffer;
	stream.avail_out = sizeof(buffer);

	inflateInit(&stream);
	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 */;
	}
	inflateEnd(&stream);
	return buf;
}

void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
{
	unsigned long mapsize;
	void *map, *buf;

	map = map_sha1_file(sha1, &mapsize);
	if (map) {
		buf = unpack_sha1_file(map, mapsize, type, size);
		munmap(map, mapsize);
		return buf;
	}
	return NULL;
}

void *read_tree_with_tree_or_commit_sha1(const unsigned char *sha1,
					 unsigned long *size,
					 unsigned char *tree_sha1_return)
{
	char type[20];
	void *buffer;
	unsigned long isize;
	int was_commit = 0;
	unsigned char tree_sha1[20];

	buffer = read_sha1_file(sha1, type, &isize);

	/* 
	 * We might have read a commit instead of a tree, in which case
	 * we parse out the tree_sha1 and attempt to read from there.
	 * (buffer + 5) is because the tree sha1 is always at offset 5
	 * in a commit record ("tree ").
	 */
	if (buffer &&
	    !strcmp(type, "commit") &&
	    !get_sha1_hex(buffer + 5, tree_sha1)) {
		free(buffer);
		buffer = read_sha1_file(tree_sha1, type, &isize);
		was_commit = 1;
	}

	/*
	 * Now do we have something and if so is it a tree?
	 */
	if (!buffer || strcmp(type, "tree")) {
		free(buffer);
		return NULL;
	}

	*size = isize;
	if (tree_sha1_return)
		memcpy(tree_sha1_return, was_commit ? tree_sha1 : sha1, 20);
	return buffer;
}

int write_sha1_file(char *buf, unsigned len, unsigned char *returnsha1)
{
	int size;
	char *compressed;
	z_stream stream;
	unsigned char sha1[20];
	SHA_CTX c;
	char *filename;
	int fd;

	/* Sha1.. */
	SHA1_Init(&c);
	SHA1_Update(&c, buf, len);
	SHA1_Final(sha1, &c);

	if (returnsha1)
		memcpy(returnsha1, sha1, 20);

	filename = sha1_file_name(sha1);
	fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
	if (fd < 0) {
		if (errno != EEXIST)
			return -1;

		/*
		 * We might do collision checking here, but we'd need to
		 * uncompress the old file and check it. Later.
		 */
		return 0;
	}

	/* 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 */;
	deflateEnd(&stream);
	size = stream.total_out;

	if (write(fd, compressed, size) != size)
		die("unable to write file");
	close(fd);
		
	return 0;
}

static inline int collision_check(char *filename, void *buf, unsigned int size)
{
#ifdef COLLISION_CHECK
	void *map;
	int fd = open(filename, O_RDONLY);
	struct stat st;
	int cmp;

	/* Unreadable object, or object went away? Strange. */
	if (fd < 0)
		return -1;

	if (fstat(fd, &st) < 0 || size != st.st_size)
		return -1;

	map = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
	close(fd);
	if (map == MAP_FAILED)
		return -1;
	cmp = memcmp(buf, map, size);
	munmap(map, size);
	if (cmp)
		return -1;
#endif
	return 0;
}

int write_sha1_buffer(const unsigned char *sha1, void *buf, unsigned int size)
{
	char *filename = sha1_file_name(sha1);
	int fd;

	fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
	if (fd < 0) {
		if (errno != EEXIST)
			return -1;
		if (collision_check(filename, buf, size))
			return error("SHA1 collision detected!"
					" This is bad, bad, BAD!\a\n");
		return 0;
	}
	write(fd, buf, size);
	close(fd);
	return 0;
}

int write_sha1_from_fd(const unsigned char *sha1, int fd)
{
	char *filename = sha1_file_name(sha1);

	int local;
	z_stream stream;
	unsigned char real_sha1[20];
	char buf[4096];
	char discard[4096];
	int ret;
	SHA_CTX c;

	local = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);

	if (local < 0)
		return error("Couldn't open %s\n", filename);

	memset(&stream, 0, sizeof(stream));

	inflateInit(&stream);

	SHA1_Init(&c);

	do {
		ssize_t size;
		size = read(fd, buf, 4096);
		if (size <= 0) {
			close(local);
			unlink(filename);
			if (!size)
				return error("Connection closed?");
			perror("Reading from connection");
			return -1;
		}
		write(local, buf, size);
		stream.avail_in = size;
		stream.next_in = buf;
		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);
		
	} while (ret == Z_OK);
	inflateEnd(&stream);

	close(local);
	SHA1_Final(real_sha1, &c);
	if (ret != Z_STREAM_END) {
		unlink(filename);
		return error("File %s corrupted", sha1_to_hex(sha1));
	}
	if (memcmp(sha1, real_sha1, 20)) {
		unlink(filename);
		return error("File %s has bad hash\n", sha1_to_hex(sha1));
	}
	
	return 0;
}

int has_sha1_file(const unsigned char *sha1)
{
	char *filename = sha1_file_name(sha1);
	struct stat st;

	if (!stat(filename, &st))
		return 1;
	return 0;
}
Commit	Line	Data
0fcfd160 LT	1	/*
	2	* GIT - The information manager from hell
	3	*
	4	* Copyright (C) Linus Torvalds, 2005
	5	*
	6	* This handles basic git sha1 object files - packing, unpacking,
	7	* creation etc.
	8	*/
	9	#include <stdarg.h>
	10	#include "cache.h"
	11
	12	const char *sha1_file_directory = NULL;
	13
144bde78 LT	14	#ifndef O_NOATIME
	15	#if defined(__linux__) && (defined(__i386__) \|\| defined(__PPC__))
	16	#define O_NOATIME 01000000
	17	#else
	18	#define O_NOATIME 0
	19	#endif
	20	#endif
	21
	22	static unsigned int sha1_file_open_flag = O_NOATIME;
	23
0fcfd160 LT	24	static unsigned hexval(char c)
	25	{
	26	if (c >= '0' && c <= '9')
	27	return c - '0';
	28	if (c >= 'a' && c <= 'f')
	29	return c - 'a' + 10;
	30	if (c >= 'A' && c <= 'F')
	31	return c - 'A' + 10;
	32	return ~0;
	33	}
	34
	35	int get_sha1_hex(const char hex, unsigned char sha1)
	36	{
	37	int i;
	38	for (i = 0; i < 20; i++) {
	39	unsigned int val = (hexval(hex[0]) << 4) \| hexval(hex[1]);
	40	if (val & ~0xff)
	41	return -1;
	42	*sha1++ = val;
	43	hex += 2;
	44	}
	45	return 0;
	46	}
	47
	48	char * sha1_to_hex(const unsigned char *sha1)
	49	{
	50	static char buffer[50];
	51	static const char hex[] = "0123456789abcdef";
	52	char *buf = buffer;
	53	int i;
	54
	55	for (i = 0; i < 20; i++) {
	56	unsigned int val = *sha1++;
	57	*buf++ = hex[val >> 4];
	58	*buf++ = hex[val & 0xf];
	59	}
	60	return buffer;
	61	}
	62
	63	/*
	64	* NOTE! This returns a statically allocated buffer, so you have to be
	65	* careful about using it. Do a "strdup()" if you need to save the
	66	* filename.
	67	*/
	68	char sha1_file_name(const unsigned char sha1)
	69	{
	70	int i;
	71	static char name, base;
	72
	73	if (!base) {
	74	char *sha1_file_directory = getenv(DB_ENVIRONMENT) ? : DEFAULT_DB_ENVIRONMENT;
	75	int len = strlen(sha1_file_directory);
	76	base = malloc(len + 60);
	77	memcpy(base, sha1_file_directory, len);
	78	memset(base+len, 0, 60);
	79	base[len] = '/';
	80	base[len+3] = '/';
	81	name = base + len + 1;
	82	}
	83	for (i = 0; i < 20; i++) {
	84	static char hex[] = "0123456789abcdef";
	85	unsigned int val = sha1[i];
	86	char pos = name + i2 + (i > 0);
	87	*pos++ = hex[val >> 4];
88	*pos = hex[val & 0xf];
89	}
90	return base;
91	}
92
d98b46f8	93	int check_sha1_signature(unsigned char sha1, void map, unsigned long size, const char *type)
0fcfd160	94	{
d98b46f8	95	char header[100];
0fcfd160 LT	96	unsigned char real_sha1[20];
	97	SHA_CTX c;
	98
	99	SHA1_Init(&c);
d98b46f8	100	SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
0fcfd160 LT	101	SHA1_Update(&c, map, size);
	102	SHA1_Final(real_sha1, &c);
	103	return memcmp(sha1, real_sha1, 20) ? -1 : 0;
	104	}
	105
	106	void map_sha1_file(const unsigned char sha1, unsigned long *size)
	107	{
	108	char *filename = sha1_file_name(sha1);
0fcfd160 LT	109	struct stat st;
0fcfd160 LT	110	void *map;
144bde78	111	int fd;
0fcfd160	112
144bde78	113	fd = open(filename, O_RDONLY \| sha1_file_open_flag);
0fcfd160	114	if (fd < 0) {
144bde78 LT	115	/* See if it works without O_NOATIME */
	116	switch (sha1_file_open_flag) {
	117	default:
	118	fd = open(filename, O_RDONLY);
	119	if (fd >= 0)
	120	break;
	121	/* Fallthrough */
	122	case 0:
	123	perror(filename);
	124	return NULL;
	125	}
	126
	127	/* If it failed once, it will probably fail again. Stop using O_NOATIME */
	128	sha1_file_open_flag = 0;
0fcfd160 LT	129	}
	130	if (fstat(fd, &st) < 0) {
	131	close(fd);
	132	return NULL;
	133	}
	134	map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
	135	close(fd);
	136	if (-1 == (int)(long)map)
	137	return NULL;
	138	*size = st.st_size;
	139	return map;
	140	}
	141
	142	void * unpack_sha1_file(void map, unsigned long mapsize, char type, unsigned long *size)
	143	{
	144	int ret, bytes;
	145	z_stream stream;
	146	char buffer[8192];
	147	char *buf;
	148
	149	/* Get the data stream */
	150	memset(&stream, 0, sizeof(stream));
	151	stream.next_in = map;
	152	stream.avail_in = mapsize;
	153	stream.next_out = buffer;
	154	stream.avail_out = sizeof(buffer);
	155
	156	inflateInit(&stream);
	157	ret = inflate(&stream, 0);
	158	if (sscanf(buffer, "%10s %lu", type, size) != 2)
	159	return NULL;
	160
	161	bytes = strlen(buffer) + 1;
	162	buf = malloc(*size);
	163	if (!buf)
	164	return NULL;
	165
	166	memcpy(buf, buffer + bytes, stream.total_out - bytes);
	167	bytes = stream.total_out - bytes;
	168	if (bytes < *size && ret == Z_OK) {
	169	stream.next_out = buf + bytes;
	170	stream.avail_out = *size - bytes;
	171	while (inflate(&stream, Z_FINISH) == Z_OK)
	172	/* nothing */;
	173	}
	174	inflateEnd(&stream);
	175	return buf;
	176	}
	177
	178	void * read_sha1_file(const unsigned char sha1, char type, unsigned long *size)
	179	{
	180	unsigned long mapsize;
	181	void map, buf;
	182
	183	map = map_sha1_file(sha1, &mapsize);
	184	if (map) {
	185	buf = unpack_sha1_file(map, mapsize, type, size);
	186	munmap(map, mapsize);
	187	return buf;
	188	}
	189	return NULL;
	190	}
	191
f4913f91 JH	192	void read_tree_with_tree_or_commit_sha1(const unsigned char sha1,
	193	unsigned long *size,
	194	unsigned char *tree_sha1_return)
	195	{
	196	char type[20];
	197	void *buffer;
	198	unsigned long isize;
	199	int was_commit = 0;
860edf7a	200	unsigned char tree_sha1[20];
f4913f91 JH	201
	202	buffer = read_sha1_file(sha1, type, &isize);
	203
	204	/*
	205	* We might have read a commit instead of a tree, in which case
	206	* we parse out the tree_sha1 and attempt to read from there.
	207	* (buffer + 5) is because the tree sha1 is always at offset 5
	208	* in a commit record ("tree ").
	209	*/
	210	if (buffer &&
	211	!strcmp(type, "commit") &&
	212	!get_sha1_hex(buffer + 5, tree_sha1)) {
	213	free(buffer);
	214	buffer = read_sha1_file(tree_sha1, type, &isize);
	215	was_commit = 1;
	216	}
	217
	218	/*
	219	* Now do we have something and if so is it a tree?
	220	*/
	221	if (!buffer \|\| strcmp(type, "tree")) {
	222	free(buffer);
860edf7a	223	return NULL;
f4913f91 JH	224	}
	225
	226	*size = isize;
	227	if (tree_sha1_return)
	228	memcpy(tree_sha1_return, was_commit ? tree_sha1 : sha1, 20);
	229	return buffer;
	230	}
	231
0fcfd160 LT	232	int write_sha1_file(char buf, unsigned len, unsigned char returnsha1)
	233	{
	234	int size;
	235	char *compressed;
	236	z_stream stream;
	237	unsigned char sha1[20];
	238	SHA_CTX c;
706bc531 LT	239	char *filename;
706bc531 LT	240	int fd;
0fcfd160	241
d98b46f8 LT	242	/* Sha1.. */
	243	SHA1_Init(&c);
	244	SHA1_Update(&c, buf, len);
	245	SHA1_Final(sha1, &c);
	246
706bc531 LT	247	if (returnsha1)
	248	memcpy(returnsha1, sha1, 20);
	249
	250	filename = sha1_file_name(sha1);
	251	fd = open(filename, O_WRONLY \| O_CREAT \| O_EXCL, 0666);
	252	if (fd < 0) {
	253	if (errno != EEXIST)
	254	return -1;
	255
	256	/*
	257	* We might do collision checking here, but we'd need to
	258	* uncompress the old file and check it. Later.
	259	*/
	260	return 0;
	261	}
	262
0fcfd160 LT	263	/* Set it up */
	264	memset(&stream, 0, sizeof(stream));
	265	deflateInit(&stream, Z_BEST_COMPRESSION);
	266	size = deflateBound(&stream, len);
	267	compressed = malloc(size);
	268
	269	/* Compress it */
	270	stream.next_in = buf;
	271	stream.avail_in = len;
	272	stream.next_out = compressed;
	273	stream.avail_out = size;
	274	while (deflate(&stream, Z_FINISH) == Z_OK)
	275	/* nothing */;
	276	deflateEnd(&stream);
	277	size = stream.total_out;
	278
706bc531 LT	279	if (write(fd, compressed, size) != size)
	280	die("unable to write file");
	281	close(fd);
	282
0fcfd160 LT	283	return 0;
	284	}
	285
	286	static inline int collision_check(char filename, void buf, unsigned int size)
	287	{
	288	#ifdef COLLISION_CHECK
	289	void *map;
	290	int fd = open(filename, O_RDONLY);
	291	struct stat st;
	292	int cmp;
	293
	294	/* Unreadable object, or object went away? Strange. */
	295	if (fd < 0)
	296	return -1;
	297
	298	if (fstat(fd, &st) < 0 \|\| size != st.st_size)
	299	return -1;
	300
	301	map = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
	302	close(fd);
	303	if (map == MAP_FAILED)
	304	return -1;
	305	cmp = memcmp(buf, map, size);
	306	munmap(map, size);
	307	if (cmp)
	308	return -1;
	309	#endif
	310	return 0;
	311	}
	312
	313	int write_sha1_buffer(const unsigned char sha1, void buf, unsigned int size)
	314	{
	315	char *filename = sha1_file_name(sha1);
	316	int fd;
	317
	318	fd = open(filename, O_WRONLY \| O_CREAT \| O_EXCL, 0666);
	319	if (fd < 0) {
	320	if (errno != EEXIST)
	321	return -1;
	322	if (collision_check(filename, buf, size))
	323	return error("SHA1 collision detected!"
	324	" This is bad, bad, BAD!\a\n");
	325	return 0;
	326	}
	327	write(fd, buf, size);
	328	close(fd);
	329	return 0;
	330	}
8237b185 DB	331
	332	int write_sha1_from_fd(const unsigned char *sha1, int fd)
	333	{
	334	char *filename = sha1_file_name(sha1);
	335
	336	int local;
	337	z_stream stream;
	338	unsigned char real_sha1[20];
	339	char buf[4096];
	340	char discard[4096];
	341	int ret;
	342	SHA_CTX c;
	343
	344	local = open(filename, O_WRONLY \| O_CREAT \| O_EXCL, 0666);
	345
	346	if (local < 0)
	347	return error("Couldn't open %s\n", filename);
	348
	349	memset(&stream, 0, sizeof(stream));
	350
	351	inflateInit(&stream);
	352
	353	SHA1_Init(&c);
	354
	355	do {
	356	ssize_t size;
	357	size = read(fd, buf, 4096);
	358	if (size <= 0) {
	359	close(local);
	360	unlink(filename);
	361	if (!size)
	362	return error("Connection closed?");
	363	perror("Reading from connection");
	364	return -1;
	365	}
	366	write(local, buf, size);
	367	stream.avail_in = size;
	368	stream.next_in = buf;
	369	do {
	370	stream.next_out = discard;
	371	stream.avail_out = sizeof(discard);
	372	ret = inflate(&stream, Z_SYNC_FLUSH);
	373	SHA1_Update(&c, discard, sizeof(discard) -
	374	stream.avail_out);
	375	} while (stream.avail_in && ret == Z_OK);
	376
	377	} while (ret == Z_OK);
	378	inflateEnd(&stream);
	379
	380	close(local);
	381	SHA1_Final(real_sha1, &c);
	382	if (ret != Z_STREAM_END) {
	383	unlink(filename);
	384	return error("File %s corrupted", sha1_to_hex(sha1));
	385	}
	386	if (memcmp(sha1, real_sha1, 20)) {
	387	unlink(filename);
	388	return error("File %s has bad hash\n", sha1_to_hex(sha1));
	389	}
	390
	391	return 0;
	392	}
	393
	394	int has_sha1_file(const unsigned char *sha1)
395	{
396	char *filename = sha1_file_name(sha1);
397	struct stat st;
398
399	if (!stat(filename, &st))
400	return 1;
401	return 0;
402	}