update-cache.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  */
   6 #include <signal.h>
   7 #include "cache.h"
   8
   9 /*
  10  * Default to not allowing changes to the list of files. The
  11  * tool doesn't actually care, but this makes it harder to add
  12  * files to the revision control by mistake by doing something
  13  * like "update-cache *" and suddenly having all the object
  14  * files be revision controlled.
  15  */
  16 static int allow_add = 0, allow_remove = 0, not_new = 0;
  17
  18 /* Three functions to allow overloaded pointer return; see linux/err.h */
  19 static inline void *ERR_PTR(long error)
  20 {
  21         return (void *) error;
  22 }
  23
  24 static inline long PTR_ERR(const void *ptr)
  25 {
  26         return (long) ptr;
  27 }
  28
  29 static inline long IS_ERR(const void *ptr)
  30 {
  31         return (unsigned long)ptr > (unsigned long)-1000L;
  32 }
  33
  34 static int index_fd(unsigned char *sha1, int fd, struct stat *st)
  35 {
  36         z_stream stream;
  37         unsigned long size = st->st_size;
  38         int max_out_bytes = size + 200;
  39         void *out = xmalloc(max_out_bytes);
  40         void *metadata = xmalloc(200);
  41         int metadata_size;
  42         void *in;
  43         SHA_CTX c;
  44
  45         in = "";
  46         if (size)
  47                 in = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
  48         close(fd);
  49         if (!out || (int)(long)in == -1)
  50                 return -1;
  51
  52         metadata_size = 1+sprintf(metadata, "blob %lu", size);
  53
  54         SHA1_Init(&c);
  55         SHA1_Update(&c, metadata, metadata_size);
  56         SHA1_Update(&c, in, size);
  57         SHA1_Final(sha1, &c);
  58
  59         memset(&stream, 0, sizeof(stream));
  60         deflateInit(&stream, Z_BEST_COMPRESSION);
  61
  62         /*
  63          * ASCII size + nul byte
  64          */
  65         stream.next_in = metadata;
  66         stream.avail_in = metadata_size;
  67         stream.next_out = out;
  68         stream.avail_out = max_out_bytes;
  69         while (deflate(&stream, 0) == Z_OK)
  70                 /* nothing */;
  71
  72         /*
  73          * File content
  74          */
  75         stream.next_in = in;
  76         stream.avail_in = size;
  77         while (deflate(&stream, Z_FINISH) == Z_OK)
  78                 /*nothing */;
  79
  80         deflateEnd(&stream);
  81
  82         return write_sha1_buffer(sha1, out, stream.total_out);
  83 }
  84
  85 /*
  86  * This only updates the "non-critical" parts of the directory
  87  * cache, ie the parts that aren't tracked by GIT, and only used
  88  * to validate the cache.
  89  */
  90 static void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
  91 {
  92         ce->ce_ctime.sec = htonl(st->st_ctime);
  93         ce->ce_mtime.sec = htonl(st->st_mtime);
  94 #ifdef NSEC
  95         ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
  96         ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
  97 #endif
  98         ce->ce_dev = htonl(st->st_dev);
  99         ce->ce_ino = htonl(st->st_ino);
 100         ce->ce_uid = htonl(st->st_uid);
 101         ce->ce_gid = htonl(st->st_gid);
 102         ce->ce_size = htonl(st->st_size);
 103 }
 104
 105 static int add_file_to_cache(char *path)
 106 {
 107         int size, namelen;
 108         struct cache_entry *ce;
 109         struct stat st;
 110         int fd;
 111
 112         fd = open(path, O_RDONLY);
 113         if (fd < 0) {
 114                 if (errno == ENOENT) {
 115                         if (allow_remove)
 116                                 return remove_file_from_cache(path);
 117                 }
 118                 return -1;
 119         }
 120         if (fstat(fd, &st) < 0) {
 121                 close(fd);
 122                 return -1;
 123         }
 124         namelen = strlen(path);
 125         size = cache_entry_size(namelen);
 126         ce = xmalloc(size);
 127         memset(ce, 0, size);
 128         memcpy(ce->name, path, namelen);
 129         fill_stat_cache_info(ce, &st);
 130         ce->ce_mode = create_ce_mode(st.st_mode);
 131         ce->ce_flags = htons(namelen);
 132
 133         if (index_fd(ce->sha1, fd, &st) < 0)
 134                 return -1;
 135
 136         return add_cache_entry(ce, allow_add);
 137 }
 138
 139 static int match_data(int fd, void *buffer, unsigned long size)
 140 {
 141         while (size) {
 142                 char compare[1024];
 143                 int ret = read(fd, compare, sizeof(compare));
 144
 145                 if (ret <= 0 || ret > size || memcmp(buffer, compare, ret))
 146                         return -1;
 147                 size -= ret;
 148                 buffer += ret;
 149         }
 150         return 0;
 151 }
 152
 153 static int compare_data(struct cache_entry *ce, unsigned long expected_size)
 154 {
 155         int match = -1;
 156         int fd = open(ce->name, O_RDONLY);
 157
 158         if (fd >= 0) {
 159                 void *buffer;
 160                 unsigned long size;
 161                 char type[10];
 162
 163                 buffer = read_sha1_file(ce->sha1, type, &size);
 164                 if (buffer) {
 165                         if (size == expected_size && !strcmp(type, "blob"))
 166                                 match = match_data(fd, buffer, size);
 167                         free(buffer);
 168                 }
 169                 close(fd);
 170         }
 171         return match;
 172 }
 173
 174 /*
 175  * "refresh" does not calculate a new sha1 file or bring the
 176  * cache up-to-date for mode/content changes. But what it
 177  * _does_ do is to "re-match" the stat information of a file
 178  * with the cache, so that you can refresh the cache for a
 179  * file that hasn't been changed but where the stat entry is
 180  * out of date.
 181  *
 182  * For example, you'd want to do this after doing a "read-tree",
 183  * to link up the stat cache details with the proper files.
 184  */
 185 static struct cache_entry *refresh_entry(struct cache_entry *ce)
 186 {
 187         struct stat st;
 188         struct cache_entry *updated;
 189         int changed, size;
 190
 191         if (stat(ce->name, &st) < 0)
 192                 return ERR_PTR(-errno);
 193
 194         changed = cache_match_stat(ce, &st);
 195         if (!changed)
 196                 return ce;
 197
 198         /*
 199          * If the mode has changed, there's no point in trying
 200          * to refresh the entry - it's not going to match
 201          */
 202         if (changed & MODE_CHANGED)
 203                 return ERR_PTR(-EINVAL);
 204
 205         if (compare_data(ce, st.st_size))
 206                 return ERR_PTR(-EINVAL);
 207
 208         size = ce_size(ce);
 209         updated = xmalloc(size);
 210         memcpy(updated, ce, size);
 211         fill_stat_cache_info(updated, &st);
 212         return updated;
 213 }
 214
 215 static void refresh_cache(void)
 216 {
 217         int i;
 218
 219         for (i = 0; i < active_nr; i++) {
 220                 struct cache_entry *ce, *new;
 221                 ce = active_cache[i];
 222                 if (ce_stage(ce)) {
 223                         printf("%s: needs merge\n", ce->name);
 224                         while ((i < active_nr) &&
 225                                ! strcmp(active_cache[i]->name, ce->name))
 226                                 i++;
 227                         i--;
 228                         continue;
 229                 }
 230
 231                 new = refresh_entry(ce);
 232                 if (IS_ERR(new)) {
 233                         if (!(not_new && PTR_ERR(new) == -ENOENT))
 234                                 printf("%s: needs update\n", ce->name);
 235                         continue;
 236                 }
 237                 active_cache[i] = new;
 238         }
 239 }
 240
 241 /*
 242  * We fundamentally don't like some paths: we don't want
 243  * dot or dot-dot anywhere, and in fact, we don't even want
 244  * any other dot-files (.git or anything else). They
 245  * are hidden, for chist sake.
 246  *
 247  * Also, we don't want double slashes or slashes at the
 248  * end that can make pathnames ambiguous.
 249  */
 250 static int verify_path(char *path)
 251 {
 252         char c;
 253
 254         goto inside;
 255         for (;;) {
 256                 if (!c)
 257                         return 1;
 258                 if (c == '/') {
 259 inside:
 260                         c = *path++;
 261                         if (c != '/' && c != '.' && c != '\0')
 262                                 continue;
 263                         return 0;
 264                 }
 265                 c = *path++;
 266         }
 267 }
 268
 269 static int add_cacheinfo(char *arg1, char *arg2, char *arg3)
 270 {
 271         int size, len;
 272         unsigned int mode;
 273         unsigned char sha1[20];
 274         struct cache_entry *ce;
 275
 276         if (sscanf(arg1, "%o", &mode) != 1)
 277                 return -1;
 278         if (get_sha1_hex(arg2, sha1))
 279                 return -1;
 280         if (!verify_path(arg3))
 281                 return -1;
 282
 283         len = strlen(arg3);
 284         size = cache_entry_size(len);
 285         ce = xmalloc(size);
 286         memset(ce, 0, size);
 287
 288         memcpy(ce->sha1, sha1, 20);
 289         memcpy(ce->name, arg3, len);
 290         ce->ce_flags = htons(len);
 291         ce->ce_mode = create_ce_mode(mode);
 292         return add_cache_entry(ce, allow_add);
 293 }
 294
 295 static const char *lockfile_name = NULL;
 296
 297 static void remove_lock_file(void)
 298 {
 299         if (lockfile_name)
 300                 unlink(lockfile_name);
 301 }
 302
 303 static void remove_lock_file_on_signal(int signo)
 304 {
 305         remove_lock_file();
 306 }
 307
 308 int main(int argc, char **argv)
 309 {
 310         int i, newfd, entries;
 311         int allow_options = 1;
 312         static char lockfile[MAXPATHLEN+1];
 313         const char *indexfile = get_index_file();
 314
 315         snprintf(lockfile, sizeof(lockfile), "%s.lock", indexfile);
 316
 317         newfd = open(lockfile, O_RDWR | O_CREAT | O_EXCL, 0600);
 318         if (newfd < 0)
 319                 die("unable to create new cachefile");
 320
 321         signal(SIGINT, remove_lock_file_on_signal);
 322         atexit(remove_lock_file);
 323         lockfile_name = lockfile;
 324
 325         entries = read_cache();
 326         if (entries < 0)
 327                 die("cache corrupted");
 328
 329         for (i = 1 ; i < argc; i++) {
 330                 char *path = argv[i];
 331
 332                 if (allow_options && *path == '-') {
 333                         if (!strcmp(path, "--")) {
 334                                 allow_options = 0;
 335                                 continue;
 336                         }
 337                         if (!strcmp(path, "--add")) {
 338                                 allow_add = 1;
 339                                 continue;
 340                         }
 341                         if (!strcmp(path, "--remove")) {
 342                                 allow_remove = 1;
 343                                 continue;
 344                         }
 345                         if (!strcmp(path, "--refresh")) {
 346                                 refresh_cache();
 347                                 continue;
 348                         }
 349                         if (!strcmp(path, "--cacheinfo")) {
 350                                 if (i+3 >= argc || add_cacheinfo(argv[i+1], argv[i+2], argv[i+3]))
 351                                         die("update-cache: --cacheinfo <mode> <sha1> <path>");
 352                                 i += 3;
 353                                 continue;
 354                         }
 355                         if (!strcmp(path, "--ignore-missing")) {
 356                                 not_new = 1;
 357                                 continue;
 358                         }
 359                         die("unknown option %s", path);
 360                 }
 361                 if (!verify_path(path)) {
 362                         fprintf(stderr, "Ignoring path %s\n", argv[i]);
 363                         continue;
 364                 }
 365                 if (add_file_to_cache(path))
 366                         die("Unable to add %s to database", path);
 367         }
 368         if (write_cache(newfd, active_cache, active_nr) || rename(lockfile, indexfile))
 369                 die("Unable to write new cachefile");
 370
 371         lockfile_name = NULL;
 372         return 0;
 373 }