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

#define _XOPEN_SOURCE 500 /* glibc2 and AIX 5.3L need this */
#define _XOPEN_SOURCE_EXTENDED 1 /* AIX 5.3L 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
a6da9395 JR	1	#define _XOPEN_SOURCE 500 /* glibc2 and AIX 5.3L need this */
a6da9395 JR	2	#define _XOPEN_SOURCE_EXTENDED 1 /* AIX 5.3L needs this */
bfac5d94	3	#include <time.h>
d98b46f8 LT	4	#include "cache.h"
	5
	6	struct entry {
	7	unsigned char old_sha1[20];
	8	unsigned char new_sha1[20];
	9	int converted;
	10	};
	11
	12	#define MAXOBJECTS (1000000)
	13
	14	static struct entry *convert[MAXOBJECTS];
	15	static int nr_convert;
	16
	17	static struct entry * convert_entry(unsigned char *sha1);
	18
	19	static struct entry insert_new(unsigned char sha1, int pos)
	20	{
812666c8	21	struct entry *new = xmalloc(sizeof(struct entry));
d98b46f8 LT	22	memset(new, 0, sizeof(*new));
	23	memcpy(new->old_sha1, sha1, 20);
	24	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	25	convert[pos] = new;
	26	nr_convert++;
	27	if (nr_convert == MAXOBJECTS)
	28	die("you're kidding me - hit maximum object limit");
	29	return new;
	30	}
	31
	32	static struct entry lookup_entry(unsigned char sha1)
	33	{
	34	int low = 0, high = nr_convert;
	35
	36	while (low < high) {
	37	int next = (low + high) / 2;
	38	struct entry *n = convert[next];
	39	int cmp = memcmp(sha1, n->old_sha1, 20);
	40	if (!cmp)
	41	return n;
	42	if (cmp < 0) {
	43	high = next;
	44	continue;
	45	}
	46	low = next+1;
	47	}
	48	return insert_new(sha1, low);
	49	}
	50
d98b46f8 LT	51	static void convert_binary_sha1(void *buffer)
	52	{
	53	struct entry *entry = convert_entry(buffer);
	54	memcpy(buffer, entry->new_sha1, 20);
	55	}
	56
	57	static void convert_ascii_sha1(void *buffer)
	58	{
	59	unsigned char sha1[20];
	60	struct entry *entry;
	61
	62	if (get_sha1_hex(buffer, sha1))
79db12e8	63	die("expected sha1, got '%s'", (char*) buffer);
d98b46f8 LT	64	entry = convert_entry(sha1);
	65	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	66	}
	67
4e81304a LT	68	static unsigned int convert_mode(unsigned int mode)
	69	{
	70	unsigned int newmode;
	71
	72	newmode = mode & S_IFMT;
	73	if (S_ISREG(mode))
	74	newmode \|= (mode & 0100) ? 0755 : 0644;
	75	return newmode;
	76	}
	77
bfac5d94	78	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	79	{
812666c8	80	char *new = xmalloc(size);
a44c9a5e	81	unsigned long newlen = 0;
bfac5d94	82	unsigned long used;
bfac5d94 LT	83
	84	used = 0;
	85	while (size) {
	86	int len = 21 + strlen(buffer);
	87	char *path = strchr(buffer, ' ');
	88	unsigned char *sha1;
	89	unsigned int mode;
	90	char slash, origpath;
	91
	92	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	93	die("bad tree conversion");
4e81304a	94	mode = convert_mode(mode);
bfac5d94 LT	95	path++;
	96	if (memcmp(path, base, baselen))
	97	break;
	98	origpath = path;
	99	path += baselen;
	100	slash = strchr(path, '/');
	101	if (!slash) {
	102	newlen += sprintf(new + newlen, "%o %s", mode, path);
	103	new[newlen++] = '\0';
	104	memcpy(new + newlen, buffer + len - 20, 20);
	105	newlen += 20;
	106
	107	used += len;
	108	size -= len;
	109	buffer += len;
	110	continue;
	111	}
	112
f220fb6b	113	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	114	new[newlen++] = 0;
	115	sha1 = (unsigned char *)(new + newlen);
	116	newlen += 20;
	117
	118	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	119
	120	used += len;
	121	size -= len;
	122	buffer += len;
	123	}
	124
a44c9a5e	125	write_sha1_file(new, newlen, "tree", result_sha1);
bfac5d94 LT	126	free(new);
	127	return used;
	128	}
	129
	130	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	131	{
	132	void *orig_buffer = buffer;
	133	unsigned long orig_size = size;
	134
d98b46f8 LT	135	while (size) {
	136	int len = 1+strlen(buffer);
	137
	138	convert_binary_sha1(buffer + len);
	139
	140	len += 20;
	141	if (len > size)
	142	die("corrupt tree object");
	143	size -= len;
	144	buffer += len;
	145	}
bfac5d94 LT	146
	147	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	148	}
	149
	150	static unsigned long parse_oldstyle_date(const char *buf)
	151	{
	152	char c, *p;
	153	char buffer[100];
	154	struct tm tm;
	155	const char *formats[] = {
	156	"%c",
	157	"%a %b %d %T",
	158	"%Z",
	159	"%Y",
	160	" %Y",
	161	NULL
	162	};
	163	/* We only ever did two timezones in the bad old format .. */
	164	const char *timezones[] = {
3f053897	165	"PDT", "PST", "CEST", NULL
bfac5d94 LT	166	};
	167	const char **fmt = formats;
	168
	169	p = buffer;
	170	while (isspace(c = *buf))
	171	buf++;
	172	while ((c = *buf++) != '\n')
	173	*p++ = c;
	174	*p++ = 0;
	175	buf = buffer;
	176	memset(&tm, 0, sizeof(tm));
	177	do {
	178	const char next = strptime(buf, fmt, &tm);
	179	if (next) {
	180	if (!*next)
	181	return mktime(&tm);
	182	buf = next;
	183	} else {
	184	const char **p = timezones;
	185	while (isspace(*buf))
	186	buf++;
	187	while (*p) {
	188	if (!memcmp(buf, p, strlen(p))) {
	189	buf += strlen(*p);
	190	break;
	191	}
	192	p++;
	193	}
	194	}
	195	fmt++;
	196	} while (buf && fmt);
	197	printf("left: %s\n", buf);
	198	return mktime(&tm);
	199	}
	200
	201	static int convert_date_line(char dst, void buf, unsigned long sp)
	202	{
	203	unsigned long size = *sp;
	204	char line = buf;
	205	char *next = strchr(line, '\n');
	206	char *date = strchr(line, '>');
	207	int len;
	208
	209	if (!next \|\| !date)
	210	die("missing or bad author/committer line %s", line);
	211	next++; date += 2;
	212
	213	*buf = next;
	214	*sp = size - (next - line);
	215
	216	len = date - line;
	217	memcpy(dst, line, len);
	218	dst += len;
	219
	220	/* Is it already in new format? */
	221	if (isdigit(*date)) {
	222	int datelen = next - date;
	223	memcpy(dst, date, datelen);
bfac5d94 LT	224	return len + datelen;
	225	}
	226
93256315 LT	227	/*
	228	* Hacky hacky: one of the sparse old-style commits does not have
	229	* any date at all, but we can fake it by using the committer date.
	230	*/
	231	if (*date == '\n' && strchr(next, '>'))
	232	date = strchr(next, '>')+2;
	233
bfac5d94	234	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	235	}
d98b46f8 LT	236
bfac5d94	237	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	238	{
812666c8	239	char *new = xmalloc(size + 100);
a44c9a5e	240	unsigned long newlen = 0;
812666c8	241
bfac5d94 LT	242	// "tree <sha1>\n"
	243	memcpy(new + newlen, buffer, 46);
	244	newlen += 46;
	245	buffer += 46;
	246	size -= 46;
	247
	248	// "parent <sha1>\n"
	249	while (!memcmp(buffer, "parent ", 7)) {
	250	memcpy(new + newlen, buffer, 48);
	251	newlen += 48;
	252	buffer += 48;
	253	size -= 48;
	254	}
	255
	256	// "author xyz <xyz> date"
	257	newlen += convert_date_line(new + newlen, &buffer, &size);
	258	// "committer xyz <xyz> date"
	259	newlen += convert_date_line(new + newlen, &buffer, &size);
	260
	261	// Rest
	262	memcpy(new + newlen, buffer, size);
	263	newlen += size;
	264
a44c9a5e	265	write_sha1_file(new, newlen, "commit", result_sha1);
bfac5d94 LT	266	free(new);
	267	}
	268
	269	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	270	{
	271	void *orig_buffer = buffer;
	272	unsigned long orig_size = size;
	273
4e81304a	274	if (memcmp(buffer, "tree ", 5))
79db12e8	275	die("Bad commit '%s'", (char*) buffer);
d98b46f8 LT	276	convert_ascii_sha1(buffer+5);
	277	buffer += 46; /* "tree " + "hex sha1" + "\n" */
	278	while (!memcmp(buffer, "parent ", 7)) {
	279	convert_ascii_sha1(buffer+7);
	280	buffer += 48;
	281	}
bfac5d94	282	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	283	}
	284
	285	static struct entry * convert_entry(unsigned char *sha1)
	286	{
	287	struct entry *entry = lookup_entry(sha1);
	288	char type[20];
	289	void buffer, data;
a44c9a5e	290	unsigned long size;
d98b46f8 LT	291
	292	if (entry->converted)
	293	return entry;
	294	data = read_sha1_file(sha1, type, &size);
	295	if (!data)
	296	die("unable to read object %s", sha1_to_hex(sha1));
	297
812666c8	298	buffer = xmalloc(size);
a44c9a5e	299	memcpy(buffer, data, size);
d98b46f8	300
bfac5d94	301	if (!strcmp(type, "blob")) {
a44c9a5e	302	write_sha1_file(buffer, size, "blob", entry->new_sha1);
bfac5d94	303	} else if (!strcmp(type, "tree"))
a44c9a5e	304	convert_tree(buffer, size, entry->new_sha1);
d98b46f8	305	else if (!strcmp(type, "commit"))
a44c9a5e	306	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	307	else
d98b46f8 LT	308	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	309	entry->converted = 1;
d98b46f8 LT	310	free(buffer);
4e81304a	311	free(data);
d98b46f8 LT	312	return entry;
	313	}
	314
	315	int main(int argc, char **argv)
	316	{
	317	unsigned char sha1[20];
	318	struct entry *entry;
	319
53228a5f JH	320	setup_git_directory();
53228a5f JH	321
3c249c95	322	if (argc != 2 \|\| get_sha1(argv[1], sha1))
215a7ad1	323	usage("git-convert-objects <sha1>");
d98b46f8 LT	324
	325	entry = convert_entry(sha1);
	326	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	327	return 0;
	328	}