update-cache.c

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