[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"
#include "blob.h"
#include "commit.h"
#include "tree.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 = xcalloc(1, sizeof(struct entry));
	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, (char *) buffer + len - 20, 20);
			newlen += 20;

			used += len;
			size -= len;
			buffer = (char *) 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 = (char *) buffer + len;
	}

	write_sha1_file(new, newlen, tree_type, 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((char *) buffer + len);

		len += 20;
		if (len > size)
			die("corrupt tree object");
		size -= len;
		buffer = (char *) 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 = (char *) buffer + 46;
	size -= 46;

	/* "parent <sha1>\n" */
	while (!memcmp(buffer, "parent ", 7)) {
		memcpy(new + newlen, buffer, 48);
		newlen += 48;
		buffer = (char *) 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_type, 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((char *) buffer + 5);
	buffer = (char *) buffer + 46;    /* "tree " + "hex sha1" + "\n" */
	while (!memcmp(buffer, "parent ", 7)) {
		convert_ascii_sha1((char *) buffer + 7);
		buffer = (char *) 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_type)) {
		write_sha1_file(buffer, size, blob_type, entry->new_sha1);
	} else if (!strcmp(type, tree_type))
		convert_tree(buffer, size, entry->new_sha1);
	else if (!strcmp(type, commit_type))
		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)
		usage("git-convert-objects <sha1>");
	if (get_sha1(argv[1], sha1))
		die("Not a valid object name %s", argv[1]);

	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	4	#include "cache.h"
8e440259 PE	5	#include "blob.h"
	6	#include "commit.h"
	7	#include "tree.h"
d98b46f8 LT	8
	9	struct entry {
	10	unsigned char old_sha1[20];
	11	unsigned char new_sha1[20];
	12	int converted;
	13	};
	14
	15	#define MAXOBJECTS (1000000)
	16
	17	static struct entry *convert[MAXOBJECTS];
	18	static int nr_convert;
	19
	20	static struct entry * convert_entry(unsigned char *sha1);
	21
	22	static struct entry insert_new(unsigned char sha1, int pos)
	23	{
90321c10	24	struct entry *new = xcalloc(1, sizeof(struct entry));
d98b46f8 LT	25	memcpy(new->old_sha1, sha1, 20);
	26	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	27	convert[pos] = new;
	28	nr_convert++;
	29	if (nr_convert == MAXOBJECTS)
	30	die("you're kidding me - hit maximum object limit");
	31	return new;
	32	}
	33
	34	static struct entry lookup_entry(unsigned char sha1)
	35	{
	36	int low = 0, high = nr_convert;
	37
	38	while (low < high) {
	39	int next = (low + high) / 2;
	40	struct entry *n = convert[next];
	41	int cmp = memcmp(sha1, n->old_sha1, 20);
	42	if (!cmp)
	43	return n;
	44	if (cmp < 0) {
	45	high = next;
	46	continue;
	47	}
	48	low = next+1;
	49	}
	50	return insert_new(sha1, low);
	51	}
	52
d98b46f8 LT	53	static void convert_binary_sha1(void *buffer)
	54	{
	55	struct entry *entry = convert_entry(buffer);
	56	memcpy(buffer, entry->new_sha1, 20);
	57	}
	58
	59	static void convert_ascii_sha1(void *buffer)
	60	{
	61	unsigned char sha1[20];
	62	struct entry *entry;
	63
	64	if (get_sha1_hex(buffer, sha1))
79db12e8	65	die("expected sha1, got '%s'", (char*) buffer);
d98b46f8 LT	66	entry = convert_entry(sha1);
	67	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	68	}
	69
4e81304a LT	70	static unsigned int convert_mode(unsigned int mode)
	71	{
	72	unsigned int newmode;
	73
	74	newmode = mode & S_IFMT;
	75	if (S_ISREG(mode))
	76	newmode \|= (mode & 0100) ? 0755 : 0644;
	77	return newmode;
	78	}
	79
bfac5d94	80	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	81	{
812666c8	82	char *new = xmalloc(size);
a44c9a5e	83	unsigned long newlen = 0;
bfac5d94	84	unsigned long used;
bfac5d94 LT	85
	86	used = 0;
	87	while (size) {
	88	int len = 21 + strlen(buffer);
	89	char *path = strchr(buffer, ' ');
	90	unsigned char *sha1;
	91	unsigned int mode;
	92	char slash, origpath;
	93
	94	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	95	die("bad tree conversion");
4e81304a	96	mode = convert_mode(mode);
bfac5d94 LT	97	path++;
	98	if (memcmp(path, base, baselen))
	99	break;
	100	origpath = path;
	101	path += baselen;
	102	slash = strchr(path, '/');
	103	if (!slash) {
	104	newlen += sprintf(new + newlen, "%o %s", mode, path);
	105	new[newlen++] = '\0';
1d7f171c	106	memcpy(new + newlen, (char *) buffer + len - 20, 20);
bfac5d94 LT	107	newlen += 20;
	108
	109	used += len;
	110	size -= len;
1d7f171c	111	buffer = (char *) buffer + len;
bfac5d94 LT	112	continue;
	113	}
	114
f220fb6b	115	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	116	new[newlen++] = 0;
	117	sha1 = (unsigned char *)(new + newlen);
	118	newlen += 20;
	119
	120	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	121
	122	used += len;
	123	size -= len;
1d7f171c	124	buffer = (char *) buffer + len;
bfac5d94 LT	125	}
bfac5d94 LT	126
8e440259	127	write_sha1_file(new, newlen, tree_type, result_sha1);
bfac5d94 LT	128	free(new);
	129	return used;
	130	}
	131
	132	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	133	{
	134	void *orig_buffer = buffer;
	135	unsigned long orig_size = size;
	136
d98b46f8 LT	137	while (size) {
	138	int len = 1+strlen(buffer);
	139
1d7f171c	140	convert_binary_sha1((char *) buffer + len);
d98b46f8 LT	141
	142	len += 20;
	143	if (len > size)
	144	die("corrupt tree object");
	145	size -= len;
1d7f171c	146	buffer = (char *) buffer + len;
d98b46f8	147	}
bfac5d94 LT	148
	149	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	150	}
	151
	152	static unsigned long parse_oldstyle_date(const char *buf)
	153	{
	154	char c, *p;
	155	char buffer[100];
	156	struct tm tm;
	157	const char *formats[] = {
	158	"%c",
	159	"%a %b %d %T",
	160	"%Z",
	161	"%Y",
	162	" %Y",
	163	NULL
	164	};
	165	/* We only ever did two timezones in the bad old format .. */
	166	const char *timezones[] = {
3f053897	167	"PDT", "PST", "CEST", NULL
bfac5d94 LT	168	};
	169	const char **fmt = formats;
	170
	171	p = buffer;
	172	while (isspace(c = *buf))
	173	buf++;
	174	while ((c = *buf++) != '\n')
	175	*p++ = c;
	176	*p++ = 0;
	177	buf = buffer;
	178	memset(&tm, 0, sizeof(tm));
	179	do {
	180	const char next = strptime(buf, fmt, &tm);
	181	if (next) {
	182	if (!*next)
	183	return mktime(&tm);
	184	buf = next;
	185	} else {
	186	const char **p = timezones;
	187	while (isspace(*buf))
	188	buf++;
	189	while (*p) {
	190	if (!memcmp(buf, p, strlen(p))) {
	191	buf += strlen(*p);
	192	break;
	193	}
	194	p++;
	195	}
	196	}
	197	fmt++;
	198	} while (buf && fmt);
	199	printf("left: %s\n", buf);
	200	return mktime(&tm);
	201	}
	202
	203	static int convert_date_line(char dst, void buf, unsigned long sp)
	204	{
	205	unsigned long size = *sp;
	206	char line = buf;
	207	char *next = strchr(line, '\n');
	208	char *date = strchr(line, '>');
	209	int len;
	210
	211	if (!next \|\| !date)
	212	die("missing or bad author/committer line %s", line);
	213	next++; date += 2;
	214
	215	*buf = next;
	216	*sp = size - (next - line);
	217
	218	len = date - line;
	219	memcpy(dst, line, len);
	220	dst += len;
	221
	222	/* Is it already in new format? */
	223	if (isdigit(*date)) {
	224	int datelen = next - date;
	225	memcpy(dst, date, datelen);
bfac5d94 LT	226	return len + datelen;
	227	}
	228
93256315 LT	229	/*
	230	* Hacky hacky: one of the sparse old-style commits does not have
	231	* any date at all, but we can fake it by using the committer date.
	232	*/
	233	if (*date == '\n' && strchr(next, '>'))
	234	date = strchr(next, '>')+2;
	235
bfac5d94	236	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	237	}
d98b46f8 LT	238
bfac5d94	239	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	240	{
812666c8	241	char *new = xmalloc(size + 100);
a44c9a5e	242	unsigned long newlen = 0;
a9486b02 PR	243
a9486b02 PR	244	/* "tree <sha1>\n" */
bfac5d94 LT	245	memcpy(new + newlen, buffer, 46);
bfac5d94 LT	246	newlen += 46;
1d7f171c	247	buffer = (char *) buffer + 46;
bfac5d94 LT	248	size -= 46;
bfac5d94 LT	249
a9486b02	250	/* "parent <sha1>\n" */
bfac5d94 LT	251	while (!memcmp(buffer, "parent ", 7)) {
	252	memcpy(new + newlen, buffer, 48);
	253	newlen += 48;
1d7f171c	254	buffer = (char *) buffer + 48;
bfac5d94 LT	255	size -= 48;
	256	}
	257
a9486b02	258	/* "author xyz <xyz> date" */
bfac5d94	259	newlen += convert_date_line(new + newlen, &buffer, &size);
a9486b02	260	/* "committer xyz <xyz> date" */
bfac5d94 LT	261	newlen += convert_date_line(new + newlen, &buffer, &size);
bfac5d94 LT	262
a9486b02	263	/* Rest */
bfac5d94 LT	264	memcpy(new + newlen, buffer, size);
	265	newlen += size;
	266
8e440259 PE	267	write_sha1_file(new, newlen, commit_type, result_sha1);
8e440259 PE	268	free(new);
bfac5d94 LT	269	}
	270
	271	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	272	{
	273	void *orig_buffer = buffer;
	274	unsigned long orig_size = size;
	275
4e81304a	276	if (memcmp(buffer, "tree ", 5))
79db12e8	277	die("Bad commit '%s'", (char*) buffer);
1d7f171c FF	278	convert_ascii_sha1((char *) buffer + 5);
1d7f171c FF	279	buffer = (char ) buffer + 46; / "tree " + "hex sha1" + "\n" */
d98b46f8	280	while (!memcmp(buffer, "parent ", 7)) {
1d7f171c FF	281	convert_ascii_sha1((char *) buffer + 7);
1d7f171c FF	282	buffer = (char *) buffer + 48;
d98b46f8	283	}
bfac5d94	284	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	285	}
	286
	287	static struct entry * convert_entry(unsigned char *sha1)
	288	{
	289	struct entry *entry = lookup_entry(sha1);
	290	char type[20];
	291	void buffer, data;
a44c9a5e	292	unsigned long size;
d98b46f8 LT	293
	294	if (entry->converted)
	295	return entry;
	296	data = read_sha1_file(sha1, type, &size);
	297	if (!data)
	298	die("unable to read object %s", sha1_to_hex(sha1));
	299
812666c8	300	buffer = xmalloc(size);
a44c9a5e	301	memcpy(buffer, data, size);
8e440259 PE	302
	303	if (!strcmp(type, blob_type)) {
	304	write_sha1_file(buffer, size, blob_type, entry->new_sha1);
	305	} else if (!strcmp(type, tree_type))
a44c9a5e	306	convert_tree(buffer, size, entry->new_sha1);
8e440259	307	else if (!strcmp(type, commit_type))
a44c9a5e	308	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	309	else
d98b46f8 LT	310	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	311	entry->converted = 1;
d98b46f8 LT	312	free(buffer);
4e81304a	313	free(data);
d98b46f8 LT	314	return entry;
	315	}
	316
	317	int main(int argc, char **argv)
	318	{
	319	unsigned char sha1[20];
	320	struct entry *entry;
	321
53228a5f JH	322	setup_git_directory();
53228a5f JH	323
31fff305	324	if (argc != 2)
215a7ad1	325	usage("git-convert-objects <sha1>");
31fff305 DL	326	if (get_sha1(argv[1], sha1))
31fff305 DL	327	die("Not a valid object name %s", argv[1]);
d98b46f8 LT	328
	329	entry = convert_entry(sha1);
	330	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	331	return 0;
	332	}