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