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