[thirdparty/git.git] / convert-objects.c

#define _XOPEN_SOURCE /* glibc2 needs this */
#include <time.h>
#include "cache.h"

struct entry {
	unsigned char old_sha1[20];
	unsigned char new_sha1[20];
	int converted;
};

#define MAXOBJECTS (1000000)

static struct entry *convert[MAXOBJECTS];
static int nr_convert;

static struct entry * convert_entry(unsigned char *sha1);

static struct entry *insert_new(unsigned char *sha1, int pos)
{
	struct entry *new = xmalloc(sizeof(struct entry));
	memset(new, 0, sizeof(*new));
	memcpy(new->old_sha1, sha1, 20);
	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	convert[pos] = new;
	nr_convert++;
	if (nr_convert == MAXOBJECTS)
		die("you're kidding me - hit maximum object limit");
	return new;
}

static struct entry *lookup_entry(unsigned char *sha1)
{
	int low = 0, high = nr_convert;

	while (low < high) {
		int next = (low + high) / 2;
		struct entry *n = convert[next];
		int cmp = memcmp(sha1, n->old_sha1, 20);
		if (!cmp)
			return n;
		if (cmp < 0) {
			high = next;
			continue;
		}
		low = next+1;
	}
	return insert_new(sha1, low);
}

static void convert_binary_sha1(void *buffer)
{
	struct entry *entry = convert_entry(buffer);
	memcpy(buffer, entry->new_sha1, 20);
}

static void convert_ascii_sha1(void *buffer)
{
	unsigned char sha1[20];
	struct entry *entry;

	if (get_sha1_hex(buffer, sha1))
		die("expected sha1, got '%s'", (char*) buffer);
	entry = convert_entry(sha1);
	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
}

static unsigned int convert_mode(unsigned int mode)
{
	unsigned int newmode;

	newmode = mode & S_IFMT;
	if (S_ISREG(mode))
		newmode |= (mode & 0100) ? 0755 : 0644;
	return newmode;
}

static int write_subdirectory(void *buffer, unsigned long size, const char *base, int baselen, unsigned char *result_sha1)
{
	char *new = xmalloc(size);
	unsigned long newlen = 0;
	unsigned long used;

	used = 0;
	while (size) {
		int len = 21 + strlen(buffer);
		char *path = strchr(buffer, ' ');
		unsigned char *sha1;
		unsigned int mode;
		char *slash, *origpath;

		if (!path || sscanf(buffer, "%o", &mode) != 1)
			die("bad tree conversion");
		mode = convert_mode(mode);
		path++;
		if (memcmp(path, base, baselen))
			break;
		origpath = path;
		path += baselen;
		slash = strchr(path, '/');
		if (!slash) {
			newlen += sprintf(new + newlen, "%o %s", mode, path);
			new[newlen++] = '\0';
			memcpy(new + newlen, buffer + len - 20, 20);
			newlen += 20;

			used += len;
			size -= len;
			buffer += len;
			continue;
		}

		newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
		new[newlen++] = 0;
		sha1 = (unsigned char *)(new + newlen);
		newlen += 20;

		len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);

		used += len;
		size -= len;
		buffer += len;
	}

	write_sha1_file(new, newlen, "tree", result_sha1);
	free(new);
	return used;
}

static void convert_tree(void *buffer, unsigned long size, unsigned char *result_sha1)
{
	void *orig_buffer = buffer;
	unsigned long orig_size = size;

	while (size) {
		int len = 1+strlen(buffer);

		convert_binary_sha1(buffer + len);

		len += 20;
		if (len > size)
			die("corrupt tree object");
		size -= len;
		buffer += len;
	}

	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
}

static unsigned long parse_oldstyle_date(const char *buf)
{
	char c, *p;
	char buffer[100];
	struct tm tm;
	const char *formats[] = {
		"%c",
		"%a %b %d %T",
		"%Z",
		"%Y",
		" %Y",
		NULL
	};
	/* We only ever did two timezones in the bad old format .. */
	const char *timezones[] = {
		"PDT", "PST", "CEST", NULL
	};
	const char **fmt = formats;

	p = buffer;
	while (isspace(c = *buf))
		buf++;
	while ((c = *buf++) != '\n')
		*p++ = c;
	*p++ = 0;
	buf = buffer;
	memset(&tm, 0, sizeof(tm));
	do {
		const char *next = strptime(buf, *fmt, &tm);
		if (next) {
			if (!*next)
				return mktime(&tm);
			buf = next;
		} else {
			const char **p = timezones;
			while (isspace(*buf))
				buf++;
			while (*p) {
				if (!memcmp(buf, *p, strlen(*p))) {
					buf += strlen(*p);
					break;
				}
				p++;
			}
		}
		fmt++;
	} while (*buf && *fmt);
	printf("left: %s\n", buf);
	return mktime(&tm);				
}

static int convert_date_line(char *dst, void **buf, unsigned long *sp)
{
	unsigned long size = *sp;
	char *line = *buf;
	char *next = strchr(line, '\n');
	char *date = strchr(line, '>');
	int len;

	if (!next || !date)
		die("missing or bad author/committer line %s", line);
	next++; date += 2;

	*buf = next;
	*sp = size - (next - line);

	len = date - line;
	memcpy(dst, line, len);
	dst += len;

	/* Is it already in new format? */
	if (isdigit(*date)) {
		int datelen = next - date;
		memcpy(dst, date, datelen);
		return len + datelen;
	}

	/*
	 * Hacky hacky: one of the sparse old-style commits does not have
	 * any date at all, but we can fake it by using the committer date.
	 */
	if (*date == '\n' && strchr(next, '>'))
		date = strchr(next, '>')+2;

	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
}

static void convert_date(void *buffer, unsigned long size, unsigned char *result_sha1)
{
	char *new = xmalloc(size + 100);
	unsigned long newlen = 0;
	
	// "tree <sha1>\n"
	memcpy(new + newlen, buffer, 46);
	newlen += 46;
	buffer += 46;
	size -= 46;

	// "parent <sha1>\n"
	while (!memcmp(buffer, "parent ", 7)) {
		memcpy(new + newlen, buffer, 48);
		newlen += 48;
		buffer += 48;
		size -= 48;
	}

	// "author xyz <xyz> date"
	newlen += convert_date_line(new + newlen, &buffer, &size);
	// "committer xyz <xyz> date"
	newlen += convert_date_line(new + newlen, &buffer, &size);

	// Rest
	memcpy(new + newlen, buffer, size);
	newlen += size;

	write_sha1_file(new, newlen, "commit", result_sha1);
	free(new);	
}

static void convert_commit(void *buffer, unsigned long size, unsigned char *result_sha1)
{
	void *orig_buffer = buffer;
	unsigned long orig_size = size;

	if (memcmp(buffer, "tree ", 5))
		die("Bad commit '%s'", (char*) buffer);
	convert_ascii_sha1(buffer+5);
	buffer += 46;    /* "tree " + "hex sha1" + "\n" */
	while (!memcmp(buffer, "parent ", 7)) {
		convert_ascii_sha1(buffer+7);
		buffer += 48;
	}
	convert_date(orig_buffer, orig_size, result_sha1);
}

static struct entry * convert_entry(unsigned char *sha1)
{
	struct entry *entry = lookup_entry(sha1);
	char type[20];
	void *buffer, *data;
	unsigned long size;

	if (entry->converted)
		return entry;
	data = read_sha1_file(sha1, type, &size);
	if (!data)
		die("unable to read object %s", sha1_to_hex(sha1));

	buffer = xmalloc(size);
	memcpy(buffer, data, size);
	
	if (!strcmp(type, "blob")) {
		write_sha1_file(buffer, size, "blob", entry->new_sha1);
	} else if (!strcmp(type, "tree"))
		convert_tree(buffer, size, entry->new_sha1);
	else if (!strcmp(type, "commit"))
		convert_commit(buffer, size, entry->new_sha1);
	else
		die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
	entry->converted = 1;
	free(buffer);
	free(data);
	return entry;
}

int main(int argc, char **argv)
{
	unsigned char sha1[20];
	struct entry *entry;

	setup_git_directory();

	if (argc != 2 || get_sha1(argv[1], sha1))
		usage("git-convert-objects <sha1>");

	entry = convert_entry(sha1);
	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	return 0;
}
Commit	Line	Data
bfac5d94 LT	1	#define _XOPEN_SOURCE /* glibc2 needs this */
bfac5d94 LT	2	#include <time.h>
d98b46f8 LT	3	#include "cache.h"
	4
	5	struct entry {
	6	unsigned char old_sha1[20];
	7	unsigned char new_sha1[20];
	8	int converted;
	9	};
	10
	11	#define MAXOBJECTS (1000000)
	12
	13	static struct entry *convert[MAXOBJECTS];
	14	static int nr_convert;
	15
	16	static struct entry * convert_entry(unsigned char *sha1);
	17
	18	static struct entry insert_new(unsigned char sha1, int pos)
	19	{
812666c8	20	struct entry *new = xmalloc(sizeof(struct entry));
d98b46f8 LT	21	memset(new, 0, sizeof(*new));
	22	memcpy(new->old_sha1, sha1, 20);
	23	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	24	convert[pos] = new;
	25	nr_convert++;
	26	if (nr_convert == MAXOBJECTS)
	27	die("you're kidding me - hit maximum object limit");
	28	return new;
	29	}
	30
	31	static struct entry lookup_entry(unsigned char sha1)
	32	{
	33	int low = 0, high = nr_convert;
	34
	35	while (low < high) {
	36	int next = (low + high) / 2;
	37	struct entry *n = convert[next];
	38	int cmp = memcmp(sha1, n->old_sha1, 20);
	39	if (!cmp)
	40	return n;
	41	if (cmp < 0) {
	42	high = next;
	43	continue;
	44	}
	45	low = next+1;
	46	}
	47	return insert_new(sha1, low);
	48	}
	49
d98b46f8 LT	50	static void convert_binary_sha1(void *buffer)
	51	{
	52	struct entry *entry = convert_entry(buffer);
	53	memcpy(buffer, entry->new_sha1, 20);
	54	}
	55
	56	static void convert_ascii_sha1(void *buffer)
	57	{
	58	unsigned char sha1[20];
	59	struct entry *entry;
	60
	61	if (get_sha1_hex(buffer, sha1))
79db12e8	62	die("expected sha1, got '%s'", (char*) buffer);
d98b46f8 LT	63	entry = convert_entry(sha1);
	64	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	65	}
	66
4e81304a LT	67	static unsigned int convert_mode(unsigned int mode)
	68	{
	69	unsigned int newmode;
	70
	71	newmode = mode & S_IFMT;
	72	if (S_ISREG(mode))
	73	newmode \|= (mode & 0100) ? 0755 : 0644;
	74	return newmode;
	75	}
	76
bfac5d94	77	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	78	{
812666c8	79	char *new = xmalloc(size);
a44c9a5e	80	unsigned long newlen = 0;
bfac5d94	81	unsigned long used;
bfac5d94 LT	82
	83	used = 0;
	84	while (size) {
	85	int len = 21 + strlen(buffer);
	86	char *path = strchr(buffer, ' ');
	87	unsigned char *sha1;
	88	unsigned int mode;
	89	char slash, origpath;
	90
	91	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	92	die("bad tree conversion");
4e81304a	93	mode = convert_mode(mode);
bfac5d94 LT	94	path++;
	95	if (memcmp(path, base, baselen))
	96	break;
	97	origpath = path;
	98	path += baselen;
	99	slash = strchr(path, '/');
	100	if (!slash) {
	101	newlen += sprintf(new + newlen, "%o %s", mode, path);
	102	new[newlen++] = '\0';
	103	memcpy(new + newlen, buffer + len - 20, 20);
	104	newlen += 20;
	105
	106	used += len;
	107	size -= len;
	108	buffer += len;
	109	continue;
	110	}
	111
f220fb6b	112	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	113	new[newlen++] = 0;
	114	sha1 = (unsigned char *)(new + newlen);
	115	newlen += 20;
	116
	117	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	118
	119	used += len;
	120	size -= len;
	121	buffer += len;
	122	}
	123
a44c9a5e	124	write_sha1_file(new, newlen, "tree", result_sha1);
bfac5d94 LT	125	free(new);
	126	return used;
	127	}
	128
	129	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	130	{
	131	void *orig_buffer = buffer;
	132	unsigned long orig_size = size;
	133
d98b46f8 LT	134	while (size) {
	135	int len = 1+strlen(buffer);
	136
	137	convert_binary_sha1(buffer + len);
	138
	139	len += 20;
	140	if (len > size)
	141	die("corrupt tree object");
	142	size -= len;
	143	buffer += len;
	144	}
bfac5d94 LT	145
	146	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	147	}
	148
	149	static unsigned long parse_oldstyle_date(const char *buf)
	150	{
	151	char c, *p;
	152	char buffer[100];
	153	struct tm tm;
	154	const char *formats[] = {
	155	"%c",
	156	"%a %b %d %T",
	157	"%Z",
	158	"%Y",
	159	" %Y",
	160	NULL
	161	};
	162	/* We only ever did two timezones in the bad old format .. */
	163	const char *timezones[] = {
3f053897	164	"PDT", "PST", "CEST", NULL
bfac5d94 LT	165	};
	166	const char **fmt = formats;
	167
	168	p = buffer;
	169	while (isspace(c = *buf))
	170	buf++;
	171	while ((c = *buf++) != '\n')
	172	*p++ = c;
	173	*p++ = 0;
	174	buf = buffer;
	175	memset(&tm, 0, sizeof(tm));
	176	do {
	177	const char next = strptime(buf, fmt, &tm);
	178	if (next) {
	179	if (!*next)
	180	return mktime(&tm);
	181	buf = next;
	182	} else {
	183	const char **p = timezones;
	184	while (isspace(*buf))
	185	buf++;
	186	while (*p) {
	187	if (!memcmp(buf, p, strlen(p))) {
	188	buf += strlen(*p);
	189	break;
	190	}
	191	p++;
	192	}
	193	}
	194	fmt++;
	195	} while (buf && fmt);
	196	printf("left: %s\n", buf);
	197	return mktime(&tm);
	198	}
	199
	200	static int convert_date_line(char dst, void buf, unsigned long sp)
	201	{
	202	unsigned long size = *sp;
	203	char line = buf;
	204	char *next = strchr(line, '\n');
	205	char *date = strchr(line, '>');
	206	int len;
	207
	208	if (!next \|\| !date)
	209	die("missing or bad author/committer line %s", line);
	210	next++; date += 2;
	211
	212	*buf = next;
	213	*sp = size - (next - line);
	214
	215	len = date - line;
	216	memcpy(dst, line, len);
	217	dst += len;
	218
	219	/* Is it already in new format? */
	220	if (isdigit(*date)) {
	221	int datelen = next - date;
	222	memcpy(dst, date, datelen);
bfac5d94 LT	223	return len + datelen;
	224	}
	225
93256315 LT	226	/*
	227	* Hacky hacky: one of the sparse old-style commits does not have
	228	* any date at all, but we can fake it by using the committer date.
	229	*/
	230	if (*date == '\n' && strchr(next, '>'))
	231	date = strchr(next, '>')+2;
	232
bfac5d94	233	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	234	}
d98b46f8 LT	235
bfac5d94	236	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	237	{
812666c8	238	char *new = xmalloc(size + 100);
a44c9a5e	239	unsigned long newlen = 0;
812666c8	240
bfac5d94 LT	241	// "tree <sha1>\n"
	242	memcpy(new + newlen, buffer, 46);
	243	newlen += 46;
	244	buffer += 46;
	245	size -= 46;
	246
	247	// "parent <sha1>\n"
	248	while (!memcmp(buffer, "parent ", 7)) {
	249	memcpy(new + newlen, buffer, 48);
	250	newlen += 48;
	251	buffer += 48;
	252	size -= 48;
	253	}
	254
	255	// "author xyz <xyz> date"
	256	newlen += convert_date_line(new + newlen, &buffer, &size);
	257	// "committer xyz <xyz> date"
	258	newlen += convert_date_line(new + newlen, &buffer, &size);
	259
	260	// Rest
	261	memcpy(new + newlen, buffer, size);
	262	newlen += size;
	263
a44c9a5e	264	write_sha1_file(new, newlen, "commit", result_sha1);
bfac5d94 LT	265	free(new);
	266	}
	267
	268	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	269	{
	270	void *orig_buffer = buffer;
	271	unsigned long orig_size = size;
	272
4e81304a	273	if (memcmp(buffer, "tree ", 5))
79db12e8	274	die("Bad commit '%s'", (char*) buffer);
d98b46f8 LT	275	convert_ascii_sha1(buffer+5);
	276	buffer += 46; /* "tree " + "hex sha1" + "\n" */
	277	while (!memcmp(buffer, "parent ", 7)) {
	278	convert_ascii_sha1(buffer+7);
	279	buffer += 48;
	280	}
bfac5d94	281	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	282	}
	283
	284	static struct entry * convert_entry(unsigned char *sha1)
	285	{
	286	struct entry *entry = lookup_entry(sha1);
	287	char type[20];
	288	void buffer, data;
a44c9a5e	289	unsigned long size;
d98b46f8 LT	290
	291	if (entry->converted)
	292	return entry;
	293	data = read_sha1_file(sha1, type, &size);
	294	if (!data)
	295	die("unable to read object %s", sha1_to_hex(sha1));
	296
812666c8	297	buffer = xmalloc(size);
a44c9a5e	298	memcpy(buffer, data, size);
d98b46f8	299
bfac5d94	300	if (!strcmp(type, "blob")) {
a44c9a5e	301	write_sha1_file(buffer, size, "blob", entry->new_sha1);
bfac5d94	302	} else if (!strcmp(type, "tree"))
a44c9a5e	303	convert_tree(buffer, size, entry->new_sha1);
d98b46f8	304	else if (!strcmp(type, "commit"))
a44c9a5e	305	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	306	else
d98b46f8 LT	307	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	308	entry->converted = 1;
d98b46f8 LT	309	free(buffer);
4e81304a	310	free(data);
d98b46f8 LT	311	return entry;
	312	}
	313
	314	int main(int argc, char **argv)
	315	{
	316	unsigned char sha1[20];
	317	struct entry *entry;
	318
53228a5f JH	319	setup_git_directory();
53228a5f JH	320
3c249c95	321	if (argc != 2 \|\| get_sha1(argv[1], sha1))
215a7ad1	322	usage("git-convert-objects <sha1>");
d98b46f8 LT	323
	324	entry = convert_entry(sha1);
	325	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	326	return 0;
	327	}