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

#define _XOPEN_SOURCE 600 /* glibc2 and AIX 5.3L need 500, OpenBSD needs 600 for S_ISLNK() */
#define _XOPEN_SOURCE_EXTENDED 1 /* AIX 5.3L needs this */
#define _GNU_SOURCE
#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));
	hashcpy(new->old_sha1, sha1);
	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 = hashcmp(sha1, n->old_sha1);
		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);
	hashcpy(buffer, entry->new_sha1);
}

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';
			hashcpy((unsigned char*)new + newlen, (unsigned char *) buffer + len - 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
73fbd33c	1	#define _XOPEN_SOURCE 600 /* glibc2 and AIX 5.3L need 500, OpenBSD needs 600 for S_ISLNK() */
a6da9395	2	#define _XOPEN_SOURCE_EXTENDED 1 /* AIX 5.3L needs this */
da7bad50	3	#define _GNU_SOURCE
bfac5d94	4	#include <time.h>
d98b46f8	5	#include "cache.h"
8e440259 PE	6	#include "blob.h"
	7	#include "commit.h"
	8	#include "tree.h"
d98b46f8 LT	9
	10	struct entry {
	11	unsigned char old_sha1[20];
	12	unsigned char new_sha1[20];
	13	int converted;
	14	};
	15
	16	#define MAXOBJECTS (1000000)
	17
	18	static struct entry *convert[MAXOBJECTS];
	19	static int nr_convert;
	20
	21	static struct entry * convert_entry(unsigned char *sha1);
	22
	23	static struct entry insert_new(unsigned char sha1, int pos)
	24	{
90321c10	25	struct entry *new = xcalloc(1, sizeof(struct entry));
e702496e	26	hashcpy(new->old_sha1, sha1);
d98b46f8 LT	27	memmove(convert + pos + 1, convert + pos, (nr_convert - pos) * sizeof(struct entry *));
	28	convert[pos] = new;
	29	nr_convert++;
	30	if (nr_convert == MAXOBJECTS)
	31	die("you're kidding me - hit maximum object limit");
	32	return new;
	33	}
	34
	35	static struct entry lookup_entry(unsigned char sha1)
	36	{
	37	int low = 0, high = nr_convert;
	38
	39	while (low < high) {
	40	int next = (low + high) / 2;
	41	struct entry *n = convert[next];
a89fccd2	42	int cmp = hashcmp(sha1, n->old_sha1);
d98b46f8 LT	43	if (!cmp)
	44	return n;
	45	if (cmp < 0) {
	46	high = next;
	47	continue;
	48	}
	49	low = next+1;
	50	}
	51	return insert_new(sha1, low);
	52	}
	53
d98b46f8 LT	54	static void convert_binary_sha1(void *buffer)
	55	{
	56	struct entry *entry = convert_entry(buffer);
e702496e	57	hashcpy(buffer, entry->new_sha1);
d98b46f8 LT	58	}
	59
	60	static void convert_ascii_sha1(void *buffer)
	61	{
	62	unsigned char sha1[20];
	63	struct entry *entry;
	64
	65	if (get_sha1_hex(buffer, sha1))
79db12e8	66	die("expected sha1, got '%s'", (char*) buffer);
d98b46f8 LT	67	entry = convert_entry(sha1);
	68	memcpy(buffer, sha1_to_hex(entry->new_sha1), 40);
	69	}
	70
4e81304a LT	71	static unsigned int convert_mode(unsigned int mode)
	72	{
	73	unsigned int newmode;
	74
	75	newmode = mode & S_IFMT;
	76	if (S_ISREG(mode))
	77	newmode \|= (mode & 0100) ? 0755 : 0644;
	78	return newmode;
	79	}
	80
bfac5d94	81	static int write_subdirectory(void buffer, unsigned long size, const char base, int baselen, unsigned char *result_sha1)
d98b46f8	82	{
812666c8	83	char *new = xmalloc(size);
a44c9a5e	84	unsigned long newlen = 0;
bfac5d94	85	unsigned long used;
bfac5d94 LT	86
	87	used = 0;
	88	while (size) {
	89	int len = 21 + strlen(buffer);
	90	char *path = strchr(buffer, ' ');
	91	unsigned char *sha1;
	92	unsigned int mode;
	93	char slash, origpath;
	94
	95	if (!path \|\| sscanf(buffer, "%o", &mode) != 1)
	96	die("bad tree conversion");
4e81304a	97	mode = convert_mode(mode);
bfac5d94 LT	98	path++;
	99	if (memcmp(path, base, baselen))
	100	break;
	101	origpath = path;
	102	path += baselen;
	103	slash = strchr(path, '/');
	104	if (!slash) {
	105	newlen += sprintf(new + newlen, "%o %s", mode, path);
	106	new[newlen++] = '\0';
e702496e	107	hashcpy((unsigned char)new + newlen, (unsigned char ) buffer + len - 20);
bfac5d94 LT	108	newlen += 20;
	109
	110	used += len;
	111	size -= len;
1d7f171c	112	buffer = (char *) buffer + len;
bfac5d94 LT	113	continue;
	114	}
	115
f220fb6b	116	newlen += sprintf(new + newlen, "%o %.*s", S_IFDIR, (int)(slash - path), path);
bfac5d94 LT	117	new[newlen++] = 0;
	118	sha1 = (unsigned char *)(new + newlen);
	119	newlen += 20;
	120
	121	len = write_subdirectory(buffer, size, origpath, slash-origpath+1, sha1);
	122
	123	used += len;
	124	size -= len;
1d7f171c	125	buffer = (char *) buffer + len;
bfac5d94 LT	126	}
bfac5d94 LT	127
8e440259	128	write_sha1_file(new, newlen, tree_type, result_sha1);
bfac5d94 LT	129	free(new);
	130	return used;
	131	}
	132
	133	static void convert_tree(void buffer, unsigned long size, unsigned char result_sha1)
	134	{
	135	void *orig_buffer = buffer;
	136	unsigned long orig_size = size;
	137
d98b46f8 LT	138	while (size) {
	139	int len = 1+strlen(buffer);
	140
1d7f171c	141	convert_binary_sha1((char *) buffer + len);
d98b46f8 LT	142
	143	len += 20;
	144	if (len > size)
	145	die("corrupt tree object");
	146	size -= len;
1d7f171c	147	buffer = (char *) buffer + len;
d98b46f8	148	}
bfac5d94 LT	149
	150	write_subdirectory(orig_buffer, orig_size, "", 0, result_sha1);
	151	}
	152
	153	static unsigned long parse_oldstyle_date(const char *buf)
	154	{
	155	char c, *p;
	156	char buffer[100];
	157	struct tm tm;
	158	const char *formats[] = {
	159	"%c",
	160	"%a %b %d %T",
	161	"%Z",
	162	"%Y",
	163	" %Y",
	164	NULL
	165	};
	166	/* We only ever did two timezones in the bad old format .. */
	167	const char *timezones[] = {
3f053897	168	"PDT", "PST", "CEST", NULL
bfac5d94 LT	169	};
	170	const char **fmt = formats;
	171
	172	p = buffer;
	173	while (isspace(c = *buf))
	174	buf++;
	175	while ((c = *buf++) != '\n')
	176	*p++ = c;
	177	*p++ = 0;
	178	buf = buffer;
	179	memset(&tm, 0, sizeof(tm));
	180	do {
	181	const char next = strptime(buf, fmt, &tm);
	182	if (next) {
	183	if (!*next)
	184	return mktime(&tm);
	185	buf = next;
	186	} else {
	187	const char **p = timezones;
	188	while (isspace(*buf))
	189	buf++;
	190	while (*p) {
	191	if (!memcmp(buf, p, strlen(p))) {
	192	buf += strlen(*p);
	193	break;
	194	}
	195	p++;
	196	}
	197	}
	198	fmt++;
	199	} while (buf && fmt);
	200	printf("left: %s\n", buf);
	201	return mktime(&tm);
	202	}
	203
	204	static int convert_date_line(char dst, void buf, unsigned long sp)
	205	{
	206	unsigned long size = *sp;
	207	char line = buf;
	208	char *next = strchr(line, '\n');
	209	char *date = strchr(line, '>');
	210	int len;
	211
	212	if (!next \|\| !date)
	213	die("missing or bad author/committer line %s", line);
	214	next++; date += 2;
	215
	216	*buf = next;
	217	*sp = size - (next - line);
	218
	219	len = date - line;
	220	memcpy(dst, line, len);
	221	dst += len;
	222
	223	/* Is it already in new format? */
	224	if (isdigit(*date)) {
	225	int datelen = next - date;
	226	memcpy(dst, date, datelen);
bfac5d94 LT	227	return len + datelen;
	228	}
	229
93256315 LT	230	/*
	231	* Hacky hacky: one of the sparse old-style commits does not have
	232	* any date at all, but we can fake it by using the committer date.
	233	*/
	234	if (*date == '\n' && strchr(next, '>'))
	235	date = strchr(next, '>')+2;
	236
bfac5d94	237	return len + sprintf(dst, "%lu -0700\n", parse_oldstyle_date(date));
d98b46f8 LT	238	}
d98b46f8 LT	239
bfac5d94	240	static void convert_date(void buffer, unsigned long size, unsigned char result_sha1)
d98b46f8	241	{
812666c8	242	char *new = xmalloc(size + 100);
a44c9a5e	243	unsigned long newlen = 0;
a9486b02 PR	244
a9486b02 PR	245	/* "tree <sha1>\n" */
bfac5d94 LT	246	memcpy(new + newlen, buffer, 46);
bfac5d94 LT	247	newlen += 46;
1d7f171c	248	buffer = (char *) buffer + 46;
bfac5d94 LT	249	size -= 46;
bfac5d94 LT	250
a9486b02	251	/* "parent <sha1>\n" */
bfac5d94 LT	252	while (!memcmp(buffer, "parent ", 7)) {
	253	memcpy(new + newlen, buffer, 48);
	254	newlen += 48;
1d7f171c	255	buffer = (char *) buffer + 48;
bfac5d94 LT	256	size -= 48;
	257	}
	258
a9486b02	259	/* "author xyz <xyz> date" */
bfac5d94	260	newlen += convert_date_line(new + newlen, &buffer, &size);
a9486b02	261	/* "committer xyz <xyz> date" */
bfac5d94 LT	262	newlen += convert_date_line(new + newlen, &buffer, &size);
bfac5d94 LT	263
a9486b02	264	/* Rest */
bfac5d94 LT	265	memcpy(new + newlen, buffer, size);
	266	newlen += size;
	267
8e440259 PE	268	write_sha1_file(new, newlen, commit_type, result_sha1);
8e440259 PE	269	free(new);
bfac5d94 LT	270	}
	271
	272	static void convert_commit(void buffer, unsigned long size, unsigned char result_sha1)
	273	{
	274	void *orig_buffer = buffer;
	275	unsigned long orig_size = size;
	276
4e81304a	277	if (memcmp(buffer, "tree ", 5))
79db12e8	278	die("Bad commit '%s'", (char*) buffer);
1d7f171c FF	279	convert_ascii_sha1((char *) buffer + 5);
1d7f171c FF	280	buffer = (char ) buffer + 46; / "tree " + "hex sha1" + "\n" */
d98b46f8	281	while (!memcmp(buffer, "parent ", 7)) {
1d7f171c FF	282	convert_ascii_sha1((char *) buffer + 7);
1d7f171c FF	283	buffer = (char *) buffer + 48;
d98b46f8	284	}
bfac5d94	285	convert_date(orig_buffer, orig_size, result_sha1);
d98b46f8 LT	286	}
	287
	288	static struct entry * convert_entry(unsigned char *sha1)
	289	{
	290	struct entry *entry = lookup_entry(sha1);
	291	char type[20];
	292	void buffer, data;
a44c9a5e	293	unsigned long size;
d98b46f8 LT	294
	295	if (entry->converted)
	296	return entry;
	297	data = read_sha1_file(sha1, type, &size);
	298	if (!data)
	299	die("unable to read object %s", sha1_to_hex(sha1));
	300
812666c8	301	buffer = xmalloc(size);
a44c9a5e	302	memcpy(buffer, data, size);
8e440259 PE	303
	304	if (!strcmp(type, blob_type)) {
	305	write_sha1_file(buffer, size, blob_type, entry->new_sha1);
	306	} else if (!strcmp(type, tree_type))
a44c9a5e	307	convert_tree(buffer, size, entry->new_sha1);
8e440259	308	else if (!strcmp(type, commit_type))
a44c9a5e	309	convert_commit(buffer, size, entry->new_sha1);
d98b46f8 LT	310	else
d98b46f8 LT	311	die("unknown object type '%s' in %s", type, sha1_to_hex(sha1));
d98b46f8 LT	312	entry->converted = 1;
d98b46f8 LT	313	free(buffer);
4e81304a	314	free(data);
d98b46f8 LT	315	return entry;
	316	}
	317
	318	int main(int argc, char **argv)
	319	{
	320	unsigned char sha1[20];
	321	struct entry *entry;
	322
53228a5f JH	323	setup_git_directory();
53228a5f JH	324
31fff305	325	if (argc != 2)
215a7ad1	326	usage("git-convert-objects <sha1>");
31fff305 DL	327	if (get_sha1(argv[1], sha1))
31fff305 DL	328	die("Not a valid object name %s", argv[1]);
d98b46f8 LT	329
	330	entry = convert_entry(sha1);
	331	printf("new sha1: %s\n", sha1_to_hex(entry->new_sha1));
	332	return 0;
	333	}