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->ctime.sec = st->st_ctime;
  72 #ifdef NSEC
  73         ce->ctime.nsec = st->st_ctim.tv_nsec;
  74 #endif
  75         ce->mtime.sec = st->st_mtime;
  76 #ifdef NSEC
  77         ce->mtime.nsec = st->st_mtim.tv_nsec;
  78 #endif
  79         ce->st_dev = st->st_dev;
  80         ce->st_ino = st->st_ino;
  81         ce->st_uid = st->st_uid;
  82         ce->st_gid = st->st_gid;
  83 }
  84
  85 static int add_file_to_cache(char *path)
  86 {
  87         int size, namelen;
  88         struct cache_entry *ce;
  89         struct stat st;
  90         int fd;
  91
  92         fd = open(path, O_RDONLY);
  93         if (fd < 0) {
  94                 if (errno == ENOENT) {
  95                         if (allow_remove)
  96                                 return remove_file_from_cache(path);
  97                 }
  98                 return -1;
  99         }
 100         if (fstat(fd, &st) < 0) {
 101                 close(fd);
 102                 return -1;
 103         }
 104         namelen = strlen(path);
 105         size = cache_entry_size(namelen);
 106         ce = malloc(size);
 107         memset(ce, 0, size);
 108         memcpy(ce->name, path, namelen);
 109         fill_stat_cache_info(ce, &st);
 110         ce->st_mode = st.st_mode;
 111         ce->st_size = st.st_size;
 112         ce->namelen = namelen;
 113
 114         if (index_fd(path, namelen, ce, fd, &st) < 0)
 115                 return -1;
 116
 117         return add_cache_entry(ce, allow_add);
 118 }
 119
 120 static int match_data(int fd, void *buffer, unsigned long size)
 121 {
 122         while (size) {
 123                 char compare[1024];
 124                 int ret = read(fd, compare, sizeof(compare));
 125
 126                 if (ret <= 0 || ret > size || memcmp(buffer, compare, ret))
 127                         return -1;
 128                 size -= ret;
 129                 buffer += ret;
 130         }
 131         return 0;
 132 }
 133
 134 static int compare_data(struct cache_entry *ce, unsigned long expected_size)
 135 {
 136         int match = -1;
 137         int fd = open(ce->name, O_RDONLY);
 138
 139         if (fd >= 0) {
 140                 void *buffer;
 141                 unsigned long size;
 142                 char type[10];
 143
 144                 buffer = read_sha1_file(ce->sha1, type, &size);
 145                 if (buffer) {
 146                         if (size == expected_size && !strcmp(type, "blob"))
 147                                 match = match_data(fd, buffer, size);
 148                         free(buffer);
 149                 }
 150                 close(fd);
 151         }
 152         return match;
 153 }
 154
 155 /*
 156  * "refresh" does not calculate a new sha1 file or bring the
 157  * cache up-to-date for mode/content changes. But what it
 158  * _does_ do is to "re-match" the stat information of a file
 159  * with the cache, so that you can refresh the cache for a
 160  * file that hasn't been changed but where the stat entry is
 161  * out of date.
 162  *
 163  * For example, you'd want to do this after doing a "read-tree",
 164  * to link up the stat cache details with the proper files.
 165  */
 166 static struct cache_entry *refresh_entry(struct cache_entry *ce)
 167 {
 168         struct stat st;
 169         struct cache_entry *updated;
 170         int changed, size;
 171
 172         if (stat(ce->name, &st) < 0)
 173                 return NULL;
 174
 175         changed = cache_match_stat(ce, &st);
 176         if (!changed)
 177                 return ce;
 178
 179         /*
 180          * If the mode has changed, there's no point in trying
 181          * to refresh the entry - it's not going to match
 182          */
 183         if (changed & MODE_CHANGED)
 184                 return NULL;
 185
 186         if (compare_data(ce, st.st_size))
 187                 return NULL;
 188
 189         size = ce_size(ce);
 190         updated = malloc(size);
 191         memcpy(updated, ce, size);
 192         fill_stat_cache_info(updated, &st);
 193         updated->st_size = st.st_size;
 194         return updated;
 195 }
 196
 197 static void refresh_cache(void)
 198 {
 199         int i;
 200
 201         for (i = 0; i < active_nr; i++) {
 202                 struct cache_entry *ce = active_cache[i];
 203                 struct cache_entry *new = refresh_entry(ce);
 204
 205                 if (!new) {
 206                         printf("%s: needs update\n", ce->name);
 207                         continue;
 208                 }
 209                 active_cache[i] = new;
 210         }
 211 }
 212
 213 /*
 214  * We fundamentally don't like some paths: we don't want
 215  * dot or dot-dot anywhere, and in fact, we don't even want
 216  * any other dot-files (.git or anything else). They
 217  * are hidden, for chist sake.
 218  *
 219  * Also, we don't want double slashes or slashes at the
 220  * end that can make pathnames ambiguous.
 221  */
 222 static int verify_path(char *path)
 223 {
 224         char c;
 225
 226         goto inside;
 227         for (;;) {
 228                 if (!c)
 229                         return 1;
 230                 if (c == '/') {
 231 inside:
 232                         c = *path++;
 233                         if (c != '/' && c != '.' && c != '\0')
 234                                 continue;
 235                         return 0;
 236                 }
 237                 c = *path++;
 238         }
 239 }
 240
 241 static int remove_lock = 0;
 242
 243 static void remove_lock_file(void)
 244 {
 245         if (remove_lock)
 246                 unlink(".git/index.lock");
 247 }
 248
 249 int main(int argc, char **argv)
 250 {
 251         int i, newfd, entries;
 252         int allow_options = 1;
 253
 254         newfd = open(".git/index.lock", O_RDWR | O_CREAT | O_EXCL, 0600);
 255         if (newfd < 0)
 256                 die("unable to create new cachefile");
 257
 258         atexit(remove_lock_file);
 259         remove_lock = 1;
 260
 261         entries = read_cache();
 262         if (entries < 0)
 263                 die("cache corrupted");
 264
 265         for (i = 1 ; i < argc; i++) {
 266                 char *path = argv[i];
 267
 268                 if (allow_options && *path == '-') {
 269                         if (!strcmp(path, "--")) {
 270                                 allow_options = 0;
 271                                 continue;
 272                         }
 273                         if (!strcmp(path, "--add")) {
 274                                 allow_add = 1;
 275                                 continue;
 276                         }
 277                         if (!strcmp(path, "--remove")) {
 278                                 allow_remove = 1;
 279                                 continue;
 280                         }
 281                         if (!strcmp(path, "--refresh")) {
 282                                 refresh_cache();
 283                                 continue;
 284                         }
 285                         die("unknown option %s", path);
 286                 }
 287                 if (!verify_path(path)) {
 288                         fprintf(stderr, "Ignoring path %s\n", argv[i]);
 289                         continue;
 290                 }
 291                 if (add_file_to_cache(path))
 292                         die("Unable to add %s to database", path);
 293         }
 294         if (write_cache(newfd, active_cache, active_nr) ||
 295             rename(".git/index.lock", ".git/index"))
 296                 die("Unable to write new cachefile");
 297
 298         remove_lock = 0;
 299         return 0;
 300 }