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