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

#define _XOPEN_SOURCE /* glibc2 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("bad sha1");
	entry = convert_entry(sha1);
	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
}

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");
		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 int convert_mode(char *buffer)
{
	char *end;
	unsigned short mode = strtoul(buffer, &end, 8);
	unsigned short newmode;
	char num[10];
	int len;

	if (*end != ' ')
		die("corrupt tree object");
	switch (mode) {
	case S_IFREG | 0644:
	case S_IFREG | 0755:
	case S_IFLNK:
	case S_IFDIR:
		return 0;
	}
	newmode = 0;
	if (S_ISREG(mode))
		newmode = (mode & 0100) ? 0755 : 0644;
	newmode |= mode & S_IFMT;
	len = sprintf(num, "%o", newmode);
	if (len != end - buffer)
		return error("unable to convert tree entry mode %o to %o", mode, newmode);
	memcpy(buffer, num, len);
	return 0;
}

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_mode(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;

	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);
	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-cache <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 */
	2	#include <time.h>
	3	#include <ctype.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))
	63	die("bad sha1");
	64	entry = convert_entry(sha1);
	65	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	66	}
	67
bfac5d94	68	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	69	{
812666c8	70	char *new = xmalloc(size);
a44c9a5e	71	unsigned long newlen = 0;
bfac5d94	72	unsigned long used;
bfac5d94 LT	73
	74	used = 0;
	75	while (size) {
	76	int len = 21 + strlen(buffer);
	77	char *path = strchr(buffer, ' ');
	78	unsigned char *sha1;
	79	unsigned int mode;
	80	char slash, origpath;
	81
	82	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	83	die("bad tree conversion");
	84	path++;
	85	if (memcmp(path, base, baselen))
	86	break;
	87	origpath = path;
	88	path += baselen;
	89	slash = strchr(path, '/');
	90	if (!slash) {
	91	newlen += sprintf(new + newlen, "%o %s", mode, path);
	92	new[newlen++] = '\0';
	93	memcpy(new + newlen, buffer + len - 20, 20);
	94	newlen += 20;
	95
	96	used += len;
	97	size -= len;
	98	buffer += len;
	99	continue;
	100	}
	101
f220fb6b	102	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	103	new[newlen++] = 0;
	104	sha1 = (unsigned char *)(new + newlen);
	105	newlen += 20;
	106
	107	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	108
	109	used += len;
	110	size -= len;
	111	buffer += len;
	112	}
	113
a44c9a5e	114	write_sha1_file(new, newlen, "tree", result_sha1);
bfac5d94 LT	115	free(new);
	116	return used;
	117	}
	118
e2418c01 LT	119	static int convert_mode(char *buffer)
	120	{
	121	char *end;
	122	unsigned short mode = strtoul(buffer, &end, 8);
	123	unsigned short newmode;
	124	char num[10];
	125	int len;
	126
	127	if (*end != ' ')
	128	die("corrupt tree object");
	129	switch (mode) {
	130	case S_IFREG \| 0644:
	131	case S_IFREG \| 0755:
	132	case S_IFLNK:
	133	case S_IFDIR:
	134	return 0;
	135	}
	136	newmode = 0;
	137	if (S_ISREG(mode))
	138	newmode = (mode & 0100) ? 0755 : 0644;
	139	newmode \|= mode & S_IFMT;
	140	len = sprintf(num, "%o", newmode);
	141	if (len != end - buffer)
	142	return error("unable to convert tree entry mode %o to %o", mode, newmode);
	143	memcpy(buffer, num, len);
	144	return 0;
	145	}
	146
bfac5d94 LT	147	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	148	{
	149	void *orig_buffer = buffer;
	150	unsigned long orig_size = size;
	151
d98b46f8 LT	152	while (size) {
	153	int len = 1+strlen(buffer);
	154
e2418c01	155	convert_mode(buffer);
d98b46f8 LT	156	convert_binary_sha1(buffer + len);
	157
	158	len += 20;
	159	if (len > size)
	160	die("corrupt tree object");
	161	size -= len;
	162	buffer += len;
	163	}
bfac5d94 LT	164
	165	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	166	}
	167
	168	static unsigned long parse_oldstyle_date(const char *buf)
	169	{
	170	char c, *p;
	171	char buffer[100];
	172	struct tm tm;
	173	const char *formats[] = {
	174	"%c",
	175	"%a %b %d %T",
	176	"%Z",
	177	"%Y",
	178	" %Y",
	179	NULL
	180	};
	181	/* We only ever did two timezones in the bad old format .. */
	182	const char *timezones[] = {
3f053897	183	"PDT", "PST", "CEST", NULL
bfac5d94 LT	184	};
	185	const char **fmt = formats;
	186
	187	p = buffer;
	188	while (isspace(c = *buf))
	189	buf++;
	190	while ((c = *buf++) != '\n')
	191	*p++ = c;
	192	*p++ = 0;
	193	buf = buffer;
	194	memset(&tm, 0, sizeof(tm));
	195	do {
	196	const char next = strptime(buf, fmt, &tm);
	197	if (next) {
	198	if (!*next)
	199	return mktime(&tm);
	200	buf = next;
	201	} else {
	202	const char **p = timezones;
	203	while (isspace(*buf))
	204	buf++;
	205	while (*p) {
	206	if (!memcmp(buf, p, strlen(p))) {
	207	buf += strlen(*p);
	208	break;
	209	}
	210	p++;
	211	}
	212	}
	213	fmt++;
	214	} while (buf && fmt);
	215	printf("left: %s\n", buf);
	216	return mktime(&tm);
	217	}
	218
	219	static int convert_date_line(char dst, void buf, unsigned long sp)
	220	{
	221	unsigned long size = *sp;
	222	char line = buf;
	223	char *next = strchr(line, '\n');
	224	char *date = strchr(line, '>');
	225	int len;
	226
	227	if (!next \|\| !date)
	228	die("missing or bad author/committer line %s", line);
	229	next++; date += 2;
	230
	231	*buf = next;
	232	*sp = size - (next - line);
	233
	234	len = date - line;
	235	memcpy(dst, line, len);
	236	dst += len;
	237
	238	/* Is it already in new format? */
	239	if (isdigit(*date)) {
	240	int datelen = next - date;
	241	memcpy(dst, date, datelen);
bfac5d94 LT	242	return len + datelen;
	243	}
	244
93256315 LT	245	/*
	246	* Hacky hacky: one of the sparse old-style commits does not have
	247	* any date at all, but we can fake it by using the committer date.
	248	*/
	249	if (*date == '\n' && strchr(next, '>'))
	250	date = strchr(next, '>')+2;
	251
bfac5d94	252	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	253	}
d98b46f8 LT	254
bfac5d94	255	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	256	{
812666c8	257	char *new = xmalloc(size + 100);
a44c9a5e	258	unsigned long newlen = 0;
812666c8	259
bfac5d94 LT	260	// "tree <sha1>\n"
	261	memcpy(new + newlen, buffer, 46);
	262	newlen += 46;
	263	buffer += 46;
	264	size -= 46;
	265
	266	// "parent <sha1>\n"
	267	while (!memcmp(buffer, "parent ", 7)) {
	268	memcpy(new + newlen, buffer, 48);
	269	newlen += 48;
	270	buffer += 48;
	271	size -= 48;
	272	}
	273
	274	// "author xyz <xyz> date"
	275	newlen += convert_date_line(new + newlen, &buffer, &size);
	276	// "committer xyz <xyz> date"
	277	newlen += convert_date_line(new + newlen, &buffer, &size);
	278
	279	// Rest
	280	memcpy(new + newlen, buffer, size);
	281	newlen += size;
	282
a44c9a5e	283	write_sha1_file(new, newlen, "commit", result_sha1);
bfac5d94 LT	284	free(new);
	285	}
	286
	287	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	288	{
	289	void *orig_buffer = buffer;
	290	unsigned long orig_size = size;
	291
d98b46f8 LT	292	convert_ascii_sha1(buffer+5);
	293	buffer += 46; /* "tree " + "hex sha1" + "\n" */
	294	while (!memcmp(buffer, "parent ", 7)) {
	295	convert_ascii_sha1(buffer+7);
	296	buffer += 48;
	297	}
bfac5d94	298	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	299	}
	300
	301	static struct entry * convert_entry(unsigned char *sha1)
	302	{
	303	struct entry *entry = lookup_entry(sha1);
	304	char type[20];
	305	void buffer, data;
a44c9a5e	306	unsigned long size;
d98b46f8 LT	307
	308	if (entry->converted)
	309	return entry;
	310	data = read_sha1_file(sha1, type, &size);
	311	if (!data)
	312	die("unable to read object %s", sha1_to_hex(sha1));
	313
812666c8	314	buffer = xmalloc(size);
a44c9a5e	315	memcpy(buffer, data, size);
d98b46f8	316
bfac5d94	317	if (!strcmp(type, "blob")) {
a44c9a5e	318	write_sha1_file(buffer, size, "blob", entry->new_sha1);
bfac5d94	319	} else if (!strcmp(type, "tree"))
a44c9a5e	320	convert_tree(buffer, size, entry->new_sha1);
d98b46f8	321	else if (!strcmp(type, "commit"))
a44c9a5e	322	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	323	else
d98b46f8 LT	324	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	325	entry->converted = 1;
	326	free(buffer);
	327	return entry;
	328	}
	329
	330	int main(int argc, char **argv)
	331	{
	332	unsigned char sha1[20];
	333	struct entry *entry;
	334
3c249c95	335	if (argc != 2 \|\| get_sha1(argv[1], sha1))
667bb59b	336	usage("git-convert-cache <sha1>");
d98b46f8 LT	337
	338	entry = convert_entry(sha1);
	339	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	340	return 0;
	341	}