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(const char *path, int namelen, struct cache_entry *ce, int fd, struct stat *st)
  18 {
  19         z_stream stream;
  20         unsigned long size = st->st_size;
  21         int max_out_bytes = namelen + size + 200;
  22         void *out = malloc(max_out_bytes);
  23         void *metadata = malloc(namelen + 200);
  24         void *in;
  25         SHA_CTX c;
  26
  27         in = "";
  28         if (size)
  29                 in = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
  30         close(fd);
  31         if (!out || (int)(long)in == -1)
  32                 return -1;
  33
  34         memset(&stream, 0, sizeof(stream));
  35         deflateInit(&stream, Z_BEST_COMPRESSION);
  36
  37         /*
  38          * ASCII size + nul byte
  39          */
  40         stream.next_in = metadata;
  41         stream.avail_in = 1+sprintf(metadata, "blob %lu", size);
  42         stream.next_out = out;
  43         stream.avail_out = max_out_bytes;
  44         while (deflate(&stream, 0) == Z_OK)
  45                 /* nothing */;
  46
  47         /*
  48          * File content
  49          */
  50         stream.next_in = in;
  51         stream.avail_in = size;
  52         while (deflate(&stream, Z_FINISH) == Z_OK)
  53                 /*nothing */;
  54
  55         deflateEnd(&stream);
  56
  57         SHA1_Init(&c);
  58         SHA1_Update(&c, out, stream.total_out);
  59         SHA1_Final(ce->sha1, &c);
  60
  61         return write_sha1_buffer(ce->sha1, out, stream.total_out);
  62 }
  63
  64 /*
  65  * This only updates the "non-critical" parts of the directory
  66  * cache, ie the parts that aren't tracked by GIT, and only used
  67  * to validate the cache.
  68  */
  69 static void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
  70 {
  71         ce->ce_ctime.sec = htonl(st->st_ctime);
  72         ce->ce_mtime.sec = htonl(st->st_mtime);
  73 #ifdef NSEC
  74         ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
  75         ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
  76 #endif
  77         ce->ce_dev = htonl(st->st_dev);
  78         ce->ce_ino = htonl(st->st_ino);
  79         ce->ce_uid = htonl(st->st_uid);
  80         ce->ce_gid = htonl(st->st_gid);
  81         ce->ce_size = htonl(st->st_size);
  82 }
  83
  84 static int add_file_to_cache(char *path)
  85 {
  86         int size, namelen;
  87         struct cache_entry *ce;
  88         struct stat st;
  89         int fd;
  90
  91         fd = open(path, O_RDONLY);
  92         if (fd < 0) {
  93                 if (errno == ENOENT) {
  94                         if (allow_remove)
  95                                 return remove_file_from_cache(path);
  96                 }
  97                 return -1;
  98         }
  99         if (fstat(fd, &st) < 0) {
 100                 close(fd);
 101                 return -1;
 102         }
 103         namelen = strlen(path);
 104         size = cache_entry_size(namelen);
 105         ce = malloc(size);
 106         memset(ce, 0, size);
 107         memcpy(ce->name, path, namelen);
 108         fill_stat_cache_info(ce, &st);
 109         ce->ce_mode = create_ce_mode(st.st_mode);
 110         ce->ce_flags = htons(namelen);
 111
 112         if (index_fd(path, namelen, ce, fd, &st) < 0)
 113                 return -1;
 114
 115         return add_cache_entry(ce, allow_add);
 116 }
 117
 118 static int match_data(int fd, void *buffer, unsigned long size)
 119 {
 120         while (size) {
 121                 char compare[1024];
 122                 int ret = read(fd, compare, sizeof(compare));
 123
 124                 if (ret <= 0 || ret > size || memcmp(buffer, compare, ret))
 125                         return -1;
 126                 size -= ret;
 127                 buffer += ret;
 128         }
 129         return 0;
 130 }
 131
 132 static int compare_data(struct cache_entry *ce, unsigned long expected_size)
 133 {
 134         int match = -1;
 135         int fd = open(ce->name, O_RDONLY);
 136
 137         if (fd >= 0) {
 138                 void *buffer;
 139                 unsigned long size;
 140                 char type[10];
 141
 142                 buffer = read_sha1_file(ce->sha1, type, &size);
 143                 if (buffer) {
 144                         if (size == expected_size && !strcmp(type, "blob"))
 145                                 match = match_data(fd, buffer, size);
 146                         free(buffer);
 147                 }
 148                 close(fd);
 149         }
 150         return match;
 151 }
 152
 153 /*
 154  * "refresh" does not calculate a new sha1 file or bring the
 155  * cache up-to-date for mode/content changes. But what it
 156  * _does_ do is to "re-match" the stat information of a file
 157  * with the cache, so that you can refresh the cache for a
 158  * file that hasn't been changed but where the stat entry is
 159  * out of date.
 160  *
 161  * For example, you'd want to do this after doing a "read-tree",
 162  * to link up the stat cache details with the proper files.
 163  */
 164 static struct cache_entry *refresh_entry(struct cache_entry *ce)
 165 {
 166         struct stat st;
 167         struct cache_entry *updated;
 168         int changed, size;
 169
 170         if (stat(ce->name, &st) < 0)
 171                 return NULL;
 172
 173         changed = cache_match_stat(ce, &st);
 174         if (!changed)
 175                 return ce;
 176
 177         /*
 178          * If the mode has changed, there's no point in trying
 179          * to refresh the entry - it's not going to match
 180          */
 181         if (changed & MODE_CHANGED)
 182                 return NULL;
 183
 184         if (compare_data(ce, st.st_size))
 185                 return NULL;
 186
 187         size = ce_size(ce);
 188         updated = malloc(size);
 189         memcpy(updated, ce, size);
 190         fill_stat_cache_info(updated, &st);
 191         return updated;
 192 }
 193
 194 static void refresh_cache(void)
 195 {
 196         int i;
 197
 198         for (i = 0; i < active_nr; i++) {
 199                 struct cache_entry *ce = active_cache[i];
 200                 struct cache_entry *new = refresh_entry(ce);
 201
 202                 if (!new) {
 203                         printf("%s: needs update\n", ce->name);
 204                         continue;
 205                 }
 206                 active_cache[i] = new;
 207         }
 208 }
 209
 210 /*
 211  * We fundamentally don't like some paths: we don't want
 212  * dot or dot-dot anywhere, and in fact, we don't even want
 213  * any other dot-files (.git or anything else). They
 214  * are hidden, for chist sake.
 215  *
 216  * Also, we don't want double slashes or slashes at the
 217  * end that can make pathnames ambiguous.
 218  */
 219 static int verify_path(char *path)
 220 {
 221         char c;
 222
 223         goto inside;
 224         for (;;) {
 225                 if (!c)
 226                         return 1;
 227                 if (c == '/') {
 228 inside:
 229                         c = *path++;
 230                         if (c != '/' && c != '.' && c != '\0')
 231                                 continue;
 232                         return 0;
 233                 }
 234                 c = *path++;
 235         }
 236 }
 237
 238 static int add_cacheinfo(char *arg1, char *arg2, char *arg3)
 239 {
 240         int size, len;
 241         unsigned int mode;
 242         unsigned char sha1[20];
 243         struct cache_entry *ce;
 244
 245         if (sscanf(arg1, "%o", &mode) != 1)
 246                 return -1;
 247         printf("got mode %o\n", mode);
 248         if (get_sha1_hex(arg2, sha1))
 249                 return -1;
 250         printf("got sha1 %s\n", sha1_to_hex(sha1));
 251         if (!verify_path(arg3))
 252                 return -1;
 253         printf("got path %s\n", arg3);
 254
 255         len = strlen(arg3);
 256         size = cache_entry_size(len);
 257         ce = malloc(size);
 258         memset(ce, 0, size);
 259
 260         memcpy(ce->sha1, sha1, 20);
 261         memcpy(ce->name, arg3, len);
 262         ce->ce_flags = htons(len);
 263         ce->ce_mode = create_ce_mode(mode);
 264         return add_cache_entry(ce, allow_add);
 265 }
 266
 267 static int remove_lock = 0;
 268
 269 static void remove_lock_file(void)
 270 {
 271         if (remove_lock)
 272                 unlink(".git/index.lock");
 273 }
 274
 275 int main(int argc, char **argv)
 276 {
 277         int i, newfd, entries;
 278         int allow_options = 1;
 279
 280         newfd = open(".git/index.lock", O_RDWR | O_CREAT | O_EXCL, 0600);
 281         if (newfd < 0)
 282                 die("unable to create new cachefile");
 283
 284         atexit(remove_lock_file);
 285         remove_lock = 1;
 286
 287         entries = read_cache();
 288         if (entries < 0)
 289                 die("cache corrupted");
 290
 291         for (i = 1 ; i < argc; i++) {
 292                 char *path = argv[i];
 293
 294                 if (allow_options && *path == '-') {
 295                         if (!strcmp(path, "--")) {
 296                                 allow_options = 0;
 297                                 continue;
 298                         }
 299                         if (!strcmp(path, "--add")) {
 300                                 allow_add = 1;
 301                                 continue;
 302                         }
 303                         if (!strcmp(path, "--remove")) {
 304                                 allow_remove = 1;
 305                                 continue;
 306                         }
 307                         if (!strcmp(path, "--refresh")) {
 308                                 refresh_cache();
 309                                 continue;
 310                         }
 311                         if (!strcmp(path, "--cacheinfo")) {
 312                                 if (i+3 >= argc || add_cacheinfo(argv[i+1], argv[i+2], argv[i+3]))
 313                                         die("update-cache: --cacheinfo <mode> <sha1> <path>");
 314                                 i += 3;
 315                                 continue;
 316                         }
 317                         die("unknown option %s", path);
 318                 }
 319                 if (!verify_path(path)) {
 320                         fprintf(stderr, "Ignoring path %s\n", argv[i]);
 321                         continue;
 322                 }
 323                 if (add_file_to_cache(path))
 324                         die("Unable to add %s to database", path);
 325         }
 326         if (write_cache(newfd, active_cache, active_nr) ||
 327             rename(".git/index.lock", ".git/index"))
 328                 die("Unable to write new cachefile");
 329
 330         remove_lock = 0;
 331         return 0;
 332 }