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

#define _XOPEN_SOURCE /* glibc2 needs this */
#define __EXTENSIONS__ /* solaris needs this */
#include <time.h>
#include <ctype.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;

	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	1	#define _XOPEN_SOURCE /* glibc2 needs this */
f0ebff0d	2	#define __EXTENSIONS__ /* solaris needs this */
bfac5d94 LT	3	#include <time.h>
bfac5d94 LT	4	#include <ctype.h>
d98b46f8 LT	5	#include "cache.h"
	6
	7	struct entry {
	8	unsigned char old_sha1[20];
	9	unsigned char new_sha1[20];
	10	int converted;
	11	};
	12
	13	#define MAXOBJECTS (1000000)
	14
	15	static struct entry *convert[MAXOBJECTS];
	16	static int nr_convert;
	17
	18	static struct entry * convert_entry(unsigned char *sha1);
	19
	20	static struct entry insert_new(unsigned char sha1, int pos)
	21	{
812666c8	22	struct entry *new = xmalloc(sizeof(struct entry));
d98b46f8 LT	23	memset(new, 0, sizeof(*new));
	24	memcpy(new->old_sha1, sha1, 20);
	25	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	26	convert[pos] = new;
	27	nr_convert++;
	28	if (nr_convert == MAXOBJECTS)
	29	die("you're kidding me - hit maximum object limit");
	30	return new;
	31	}
	32
	33	static struct entry lookup_entry(unsigned char sha1)
	34	{
	35	int low = 0, high = nr_convert;
	36
	37	while (low < high) {
	38	int next = (low + high) / 2;
	39	struct entry *n = convert[next];
	40	int cmp = memcmp(sha1, n->old_sha1, 20);
	41	if (!cmp)
	42	return n;
	43	if (cmp < 0) {
	44	high = next;
	45	continue;
	46	}
	47	low = next+1;
	48	}
	49	return insert_new(sha1, low);
	50	}
	51
d98b46f8 LT	52	static void convert_binary_sha1(void *buffer)
	53	{
	54	struct entry *entry = convert_entry(buffer);
	55	memcpy(buffer, entry->new_sha1, 20);
	56	}
	57
	58	static void convert_ascii_sha1(void *buffer)
	59	{
	60	unsigned char sha1[20];
	61	struct entry *entry;
	62
	63	if (get_sha1_hex(buffer, sha1))
79db12e8	64	die("expected sha1, got '%s'", (char*) buffer);
d98b46f8 LT	65	entry = convert_entry(sha1);
	66	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	67	}
	68
4e81304a LT	69	static unsigned int convert_mode(unsigned int mode)
	70	{
	71	unsigned int newmode;
	72
	73	newmode = mode & S_IFMT;
	74	if (S_ISREG(mode))
	75	newmode \|= (mode & 0100) ? 0755 : 0644;
	76	return newmode;
	77	}
	78
bfac5d94	79	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	80	{
812666c8	81	char *new = xmalloc(size);
a44c9a5e	82	unsigned long newlen = 0;
bfac5d94	83	unsigned long used;
bfac5d94 LT	84
	85	used = 0;
	86	while (size) {
	87	int len = 21 + strlen(buffer);
	88	char *path = strchr(buffer, ' ');
	89	unsigned char *sha1;
	90	unsigned int mode;
	91	char slash, origpath;
	92
	93	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	94	die("bad tree conversion");
4e81304a	95	mode = convert_mode(mode);
bfac5d94 LT	96	path++;
	97	if (memcmp(path, base, baselen))
	98	break;
	99	origpath = path;
	100	path += baselen;
	101	slash = strchr(path, '/');
	102	if (!slash) {
	103	newlen += sprintf(new + newlen, "%o %s", mode, path);
	104	new[newlen++] = '\0';
	105	memcpy(new + newlen, buffer + len - 20, 20);
	106	newlen += 20;
	107
	108	used += len;
	109	size -= len;
	110	buffer += len;
	111	continue;
	112	}
	113
f220fb6b	114	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	115	new[newlen++] = 0;
	116	sha1 = (unsigned char *)(new + newlen);
	117	newlen += 20;
	118
	119	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	120
	121	used += len;
	122	size -= len;
	123	buffer += len;
	124	}
	125
a44c9a5e	126	write_sha1_file(new, newlen, "tree", result_sha1);
bfac5d94 LT	127	free(new);
	128	return used;
	129	}
	130
	131	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	132	{
	133	void *orig_buffer = buffer;
	134	unsigned long orig_size = size;
	135
d98b46f8 LT	136	while (size) {
	137	int len = 1+strlen(buffer);
	138
	139	convert_binary_sha1(buffer + len);
	140
	141	len += 20;
	142	if (len > size)
	143	die("corrupt tree object");
	144	size -= len;
	145	buffer += len;
	146	}
bfac5d94 LT	147
	148	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	149	}
	150
	151	static unsigned long parse_oldstyle_date(const char *buf)
	152	{
	153	char c, *p;
	154	char buffer[100];
	155	struct tm tm;
	156	const char *formats[] = {
	157	"%c",
	158	"%a %b %d %T",
	159	"%Z",
	160	"%Y",
	161	" %Y",
	162	NULL
	163	};
	164	/* We only ever did two timezones in the bad old format .. */
	165	const char *timezones[] = {
3f053897	166	"PDT", "PST", "CEST", NULL
bfac5d94 LT	167	};
	168	const char **fmt = formats;
	169
	170	p = buffer;
	171	while (isspace(c = *buf))
	172	buf++;
	173	while ((c = *buf++) != '\n')
	174	*p++ = c;
	175	*p++ = 0;
	176	buf = buffer;
	177	memset(&tm, 0, sizeof(tm));
	178	do {
	179	const char next = strptime(buf, fmt, &tm);
	180	if (next) {
	181	if (!*next)
	182	return mktime(&tm);
	183	buf = next;
	184	} else {
	185	const char **p = timezones;
	186	while (isspace(*buf))
	187	buf++;
	188	while (*p) {
	189	if (!memcmp(buf, p, strlen(p))) {
	190	buf += strlen(*p);
	191	break;
	192	}
	193	p++;
	194	}
	195	}
	196	fmt++;
	197	} while (buf && fmt);
	198	printf("left: %s\n", buf);
	199	return mktime(&tm);
	200	}
	201
	202	static int convert_date_line(char dst, void buf, unsigned long sp)
	203	{
	204	unsigned long size = *sp;
	205	char line = buf;
	206	char *next = strchr(line, '\n');
	207	char *date = strchr(line, '>');
	208	int len;
	209
	210	if (!next \|\| !date)
	211	die("missing or bad author/committer line %s", line);
	212	next++; date += 2;
	213
	214	*buf = next;
	215	*sp = size - (next - line);
	216
	217	len = date - line;
	218	memcpy(dst, line, len);
	219	dst += len;
	220
	221	/* Is it already in new format? */
	222	if (isdigit(*date)) {
	223	int datelen = next - date;
	224	memcpy(dst, date, datelen);
bfac5d94 LT	225	return len + datelen;
	226	}
	227
93256315 LT	228	/*
	229	* Hacky hacky: one of the sparse old-style commits does not have
	230	* any date at all, but we can fake it by using the committer date.
	231	*/
	232	if (*date == '\n' && strchr(next, '>'))
	233	date = strchr(next, '>')+2;
	234
bfac5d94	235	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	236	}
d98b46f8 LT	237
bfac5d94	238	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	239	{
812666c8	240	char *new = xmalloc(size + 100);
a44c9a5e	241	unsigned long newlen = 0;
812666c8	242
bfac5d94 LT	243	// "tree <sha1>\n"
	244	memcpy(new + newlen, buffer, 46);
	245	newlen += 46;
	246	buffer += 46;
	247	size -= 46;
	248
	249	// "parent <sha1>\n"
	250	while (!memcmp(buffer, "parent ", 7)) {
	251	memcpy(new + newlen, buffer, 48);
	252	newlen += 48;
	253	buffer += 48;
	254	size -= 48;
	255	}
	256
	257	// "author xyz <xyz> date"
	258	newlen += convert_date_line(new + newlen, &buffer, &size);
	259	// "committer xyz <xyz> date"
	260	newlen += convert_date_line(new + newlen, &buffer, &size);
	261
	262	// Rest
	263	memcpy(new + newlen, buffer, size);
	264	newlen += size;
	265
a44c9a5e	266	write_sha1_file(new, newlen, "commit", result_sha1);
bfac5d94 LT	267	free(new);
	268	}
	269
	270	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	271	{
	272	void *orig_buffer = buffer;
	273	unsigned long orig_size = size;
	274
4e81304a	275	if (memcmp(buffer, "tree ", 5))
79db12e8	276	die("Bad commit '%s'", (char*) buffer);
d98b46f8 LT	277	convert_ascii_sha1(buffer+5);
	278	buffer += 46; /* "tree " + "hex sha1" + "\n" */
	279	while (!memcmp(buffer, "parent ", 7)) {
	280	convert_ascii_sha1(buffer+7);
	281	buffer += 48;
	282	}
bfac5d94	283	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	284	}
	285
	286	static struct entry * convert_entry(unsigned char *sha1)
	287	{
	288	struct entry *entry = lookup_entry(sha1);
	289	char type[20];
	290	void buffer, data;
a44c9a5e	291	unsigned long size;
d98b46f8 LT	292
	293	if (entry->converted)
	294	return entry;
	295	data = read_sha1_file(sha1, type, &size);
	296	if (!data)
	297	die("unable to read object %s", sha1_to_hex(sha1));
	298
812666c8	299	buffer = xmalloc(size);
a44c9a5e	300	memcpy(buffer, data, size);
d98b46f8	301
bfac5d94	302	if (!strcmp(type, "blob")) {
a44c9a5e	303	write_sha1_file(buffer, size, "blob", entry->new_sha1);
bfac5d94	304	} else if (!strcmp(type, "tree"))
a44c9a5e	305	convert_tree(buffer, size, entry->new_sha1);
d98b46f8	306	else if (!strcmp(type, "commit"))
a44c9a5e	307	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	308	else
d98b46f8 LT	309	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	310	entry->converted = 1;
d98b46f8 LT	311	free(buffer);
4e81304a	312	free(data);
d98b46f8 LT	313	return entry;
	314	}
	315
	316	int main(int argc, char **argv)
	317	{
	318	unsigned char sha1[20];
	319	struct entry *entry;
	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	}