builtin/cat-file.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  */
   6 #include "cache.h"
   7 #include "builtin.h"
   8 #include "parse-options.h"
   9 #include "userdiff.h"
  10 #include "streaming.h"
  11 #include "tree-walk.h"
  12
  13 static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
  14                         int unknown_type)
  15 {
  16         unsigned char sha1[20];
  17         enum object_type type;
  18         char *buf;
  19         unsigned long size;
  20         struct object_context obj_context;
  21         struct object_info oi = {NULL};
  22         struct strbuf sb = STRBUF_INIT;
  23         unsigned flags = LOOKUP_REPLACE_OBJECT;
  24
  25         if (unknown_type)
  26                 flags |= LOOKUP_UNKNOWN_OBJECT;
  27
  28         if (get_sha1_with_context(obj_name, 0, sha1, &obj_context))
  29                 die("Not a valid object name %s", obj_name);
  30
  31         buf = NULL;
  32         switch (opt) {
  33         case 't':
  34                 oi.typename = &sb;
  35                 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
  36                         die("git cat-file: could not get object info");
  37                 if (sb.len) {
  38                         printf("%s\n", sb.buf);
  39                         strbuf_release(&sb);
  40                         return 0;
  41                 }
  42                 break;
  43
  44         case 's':
  45                 oi.sizep = &size;
  46                 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
  47                         die("git cat-file: could not get object info");
  48                 printf("%lu\n", size);
  49                 return 0;
  50
  51         case 'e':
  52                 return !has_sha1_file(sha1);
  53
  54         case 'c':
  55                 if (!obj_context.path[0])
  56                         die("git cat-file --textconv %s: <object> must be <sha1:path>",
  57                             obj_name);
  58
  59                 if (textconv_object(obj_context.path, obj_context.mode, sha1, 1, &buf, &size))
  60                         break;
  61
  62         case 'p':
  63                 type = sha1_object_info(sha1, NULL);
  64                 if (type < 0)
  65                         die("Not a valid object name %s", obj_name);
  66
  67                 /* custom pretty-print here */
  68                 if (type == OBJ_TREE) {
  69                         const char *ls_args[3] = { NULL };
  70                         ls_args[0] =  "ls-tree";
  71                         ls_args[1] =  obj_name;
  72                         return cmd_ls_tree(2, ls_args, NULL);
  73                 }
  74
  75                 if (type == OBJ_BLOB)
  76                         return stream_blob_to_fd(1, sha1, NULL, 0);
  77                 buf = read_sha1_file(sha1, &type, &size);
  78                 if (!buf)
  79                         die("Cannot read object %s", obj_name);
  80
  81                 /* otherwise just spit out the data */
  82                 break;
  83
  84         case 0:
  85                 if (type_from_string(exp_type) == OBJ_BLOB) {
  86                         unsigned char blob_sha1[20];
  87                         if (sha1_object_info(sha1, NULL) == OBJ_TAG) {
  88                                 char *buffer = read_sha1_file(sha1, &type, &size);
  89                                 const char *target;
  90                                 if (!skip_prefix(buffer, "object ", &target) ||
  91                                     get_sha1_hex(target, blob_sha1))
  92                                         die("%s not a valid tag", sha1_to_hex(sha1));
  93                                 free(buffer);
  94                         } else
  95                                 hashcpy(blob_sha1, sha1);
  96
  97                         if (sha1_object_info(blob_sha1, NULL) == OBJ_BLOB)
  98                                 return stream_blob_to_fd(1, blob_sha1, NULL, 0);
  99                         /*
 100                          * we attempted to dereference a tag to a blob
 101                          * and failed; there may be new dereference
 102                          * mechanisms this code is not aware of.
 103                          * fall-back to the usual case.
 104                          */
 105                 }
 106                 buf = read_object_with_reference(sha1, exp_type, &size, NULL);
 107                 break;
 108
 109         default:
 110                 die("git cat-file: unknown option: %s", exp_type);
 111         }
 112
 113         if (!buf)
 114                 die("git cat-file %s: bad file", obj_name);
 115
 116         write_or_die(1, buf, size);
 117         return 0;
 118 }
 119
 120 struct expand_data {
 121         unsigned char sha1[20];
 122         enum object_type type;
 123         unsigned long size;
 124         unsigned long disk_size;
 125         const char *rest;
 126         unsigned char delta_base_sha1[20];
 127
 128         /*
 129          * If mark_query is true, we do not expand anything, but rather
 130          * just mark the object_info with items we wish to query.
 131          */
 132         int mark_query;
 133
 134         /*
 135          * Whether to split the input on whitespace before feeding it to
 136          * get_sha1; this is decided during the mark_query phase based on
 137          * whether we have a %(rest) token in our format.
 138          */
 139         int split_on_whitespace;
 140
 141         /*
 142          * After a mark_query run, this object_info is set up to be
 143          * passed to sha1_object_info_extended. It will point to the data
 144          * elements above, so you can retrieve the response from there.
 145          */
 146         struct object_info info;
 147 };
 148
 149 static int is_atom(const char *atom, const char *s, int slen)
 150 {
 151         int alen = strlen(atom);
 152         return alen == slen && !memcmp(atom, s, alen);
 153 }
 154
 155 static void expand_atom(struct strbuf *sb, const char *atom, int len,
 156                         void *vdata)
 157 {
 158         struct expand_data *data = vdata;
 159
 160         if (is_atom("objectname", atom, len)) {
 161                 if (!data->mark_query)
 162                         strbuf_addstr(sb, sha1_to_hex(data->sha1));
 163         } else if (is_atom("objecttype", atom, len)) {
 164                 if (data->mark_query)
 165                         data->info.typep = &data->type;
 166                 else
 167                         strbuf_addstr(sb, typename(data->type));
 168         } else if (is_atom("objectsize", atom, len)) {
 169                 if (data->mark_query)
 170                         data->info.sizep = &data->size;
 171                 else
 172                         strbuf_addf(sb, "%lu", data->size);
 173         } else if (is_atom("objectsize:disk", atom, len)) {
 174                 if (data->mark_query)
 175                         data->info.disk_sizep = &data->disk_size;
 176                 else
 177                         strbuf_addf(sb, "%lu", data->disk_size);
 178         } else if (is_atom("rest", atom, len)) {
 179                 if (data->mark_query)
 180                         data->split_on_whitespace = 1;
 181                 else if (data->rest)
 182                         strbuf_addstr(sb, data->rest);
 183         } else if (is_atom("deltabase", atom, len)) {
 184                 if (data->mark_query)
 185                         data->info.delta_base_sha1 = data->delta_base_sha1;
 186                 else
 187                         strbuf_addstr(sb, sha1_to_hex(data->delta_base_sha1));
 188         } else
 189                 die("unknown format element: %.*s", len, atom);
 190 }
 191
 192 static size_t expand_format(struct strbuf *sb, const char *start, void *data)
 193 {
 194         const char *end;
 195
 196         if (*start != '(')
 197                 return 0;
 198         end = strchr(start + 1, ')');
 199         if (!end)
 200                 die("format element '%s' does not end in ')'", start);
 201
 202         expand_atom(sb, start + 1, end - start - 1, data);
 203
 204         return end - start + 1;
 205 }
 206
 207 static void print_object_or_die(int fd, struct expand_data *data)
 208 {
 209         const unsigned char *sha1 = data->sha1;
 210
 211         assert(data->info.typep);
 212
 213         if (data->type == OBJ_BLOB) {
 214                 if (stream_blob_to_fd(fd, sha1, NULL, 0) < 0)
 215                         die("unable to stream %s to stdout", sha1_to_hex(sha1));
 216         }
 217         else {
 218                 enum object_type type;
 219                 unsigned long size;
 220                 void *contents;
 221
 222                 contents = read_sha1_file(sha1, &type, &size);
 223                 if (!contents)
 224                         die("object %s disappeared", sha1_to_hex(sha1));
 225                 if (type != data->type)
 226                         die("object %s changed type!?", sha1_to_hex(sha1));
 227                 if (data->info.sizep && size != data->size)
 228                         die("object %s changed size!?", sha1_to_hex(sha1));
 229
 230                 write_or_die(fd, contents, size);
 231                 free(contents);
 232         }
 233 }
 234
 235 struct batch_options {
 236         int enabled;
 237         int follow_symlinks;
 238         int print_contents;
 239         const char *format;
 240 };
 241
 242 static int batch_one_object(const char *obj_name, struct batch_options *opt,
 243                             struct expand_data *data)
 244 {
 245         struct strbuf buf = STRBUF_INIT;
 246         struct object_context ctx;
 247         int flags = opt->follow_symlinks ? GET_SHA1_FOLLOW_SYMLINKS : 0;
 248         enum follow_symlinks_result result;
 249
 250         if (!obj_name)
 251            return 1;
 252
 253         result = get_sha1_with_context(obj_name, flags, data->sha1, &ctx);
 254         if (result != FOUND) {
 255                 switch (result) {
 256                 case MISSING_OBJECT:
 257                         printf("%s missing\n", obj_name);
 258                         break;
 259                 case DANGLING_SYMLINK:
 260                         printf("dangling %"PRIuMAX"\n%s\n",
 261                                (uintmax_t)strlen(obj_name), obj_name);
 262                         break;
 263                 case SYMLINK_LOOP:
 264                         printf("loop %"PRIuMAX"\n%s\n",
 265                                (uintmax_t)strlen(obj_name), obj_name);
 266                         break;
 267                 case NOT_DIR:
 268                         printf("notdir %"PRIuMAX"\n%s\n",
 269                                (uintmax_t)strlen(obj_name), obj_name);
 270                         break;
 271                 default:
 272                         die("BUG: unknown get_sha1_with_context result %d\n",
 273                                result);
 274                         break;
 275                 }
 276                 fflush(stdout);
 277                 return 0;
 278         }
 279
 280         if (ctx.mode == 0) {
 281                 printf("symlink %"PRIuMAX"\n%s\n",
 282                        (uintmax_t)ctx.symlink_path.len,
 283                        ctx.symlink_path.buf);
 284                 fflush(stdout);
 285                 return 0;
 286         }
 287
 288         if (sha1_object_info_extended(data->sha1, &data->info, LOOKUP_REPLACE_OBJECT) < 0) {
 289                 printf("%s missing\n", obj_name);
 290                 fflush(stdout);
 291                 return 0;
 292         }
 293
 294         strbuf_expand(&buf, opt->format, expand_format, data);
 295         strbuf_addch(&buf, '\n');
 296         write_or_die(1, buf.buf, buf.len);
 297         strbuf_release(&buf);
 298
 299         if (opt->print_contents) {
 300                 print_object_or_die(1, data);
 301                 write_or_die(1, "\n", 1);
 302         }
 303         return 0;
 304 }
 305
 306 static int batch_objects(struct batch_options *opt)
 307 {
 308         struct strbuf buf = STRBUF_INIT;
 309         struct expand_data data;
 310         int save_warning;
 311         int retval = 0;
 312
 313         if (!opt->format)
 314                 opt->format = "%(objectname) %(objecttype) %(objectsize)";
 315
 316         /*
 317          * Expand once with our special mark_query flag, which will prime the
 318          * object_info to be handed to sha1_object_info_extended for each
 319          * object.
 320          */
 321         memset(&data, 0, sizeof(data));
 322         data.mark_query = 1;
 323         strbuf_expand(&buf, opt->format, expand_format, &data);
 324         data.mark_query = 0;
 325
 326         /*
 327          * If we are printing out the object, then always fill in the type,
 328          * since we will want to decide whether or not to stream.
 329          */
 330         if (opt->print_contents)
 331                 data.info.typep = &data.type;
 332
 333         /*
 334          * We are going to call get_sha1 on a potentially very large number of
 335          * objects. In most large cases, these will be actual object sha1s. The
 336          * cost to double-check that each one is not also a ref (just so we can
 337          * warn) ends up dwarfing the actual cost of the object lookups
 338          * themselves. We can work around it by just turning off the warning.
 339          */
 340         save_warning = warn_on_object_refname_ambiguity;
 341         warn_on_object_refname_ambiguity = 0;
 342
 343         while (strbuf_getline(&buf, stdin, '\n') != EOF) {
 344                 if (data.split_on_whitespace) {
 345                         /*
 346                          * Split at first whitespace, tying off the beginning
 347                          * of the string and saving the remainder (or NULL) in
 348                          * data.rest.
 349                          */
 350                         char *p = strpbrk(buf.buf, " \t");
 351                         if (p) {
 352                                 while (*p && strchr(" \t", *p))
 353                                         *p++ = '\0';
 354                         }
 355                         data.rest = p;
 356                 }
 357
 358                 retval = batch_one_object(buf.buf, opt, &data);
 359                 if (retval)
 360                         break;
 361         }
 362
 363         strbuf_release(&buf);
 364         warn_on_object_refname_ambiguity = save_warning;
 365         return retval;
 366 }
 367
 368 static const char * const cat_file_usage[] = {
 369         N_("git cat-file (-t [--allow-unknown-type]|-s [--allow-unknown-type]|-e|-p|<type>|--textconv) <object>"),
 370         N_("git cat-file (--batch | --batch-check) [--follow-symlinks] < <list-of-objects>"),
 371         NULL
 372 };
 373
 374 static int git_cat_file_config(const char *var, const char *value, void *cb)
 375 {
 376         if (userdiff_config(var, value) < 0)
 377                 return -1;
 378
 379         return git_default_config(var, value, cb);
 380 }
 381
 382 static int batch_option_callback(const struct option *opt,
 383                                  const char *arg,
 384                                  int unset)
 385 {
 386         struct batch_options *bo = opt->value;
 387
 388         if (bo->enabled) {
 389                 return 1;
 390         }
 391
 392         bo->enabled = 1;
 393         bo->print_contents = !strcmp(opt->long_name, "batch");
 394         bo->format = arg;
 395
 396         return 0;
 397 }
 398
 399 int cmd_cat_file(int argc, const char **argv, const char *prefix)
 400 {
 401         int opt = 0;
 402         const char *exp_type = NULL, *obj_name = NULL;
 403         struct batch_options batch = {0};
 404         int unknown_type = 0;
 405
 406         const struct option options[] = {
 407                 OPT_GROUP(N_("<type> can be one of: blob, tree, commit, tag")),
 408                 OPT_CMDMODE('t', NULL, &opt, N_("show object type"), 't'),
 409                 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
 410                 OPT_CMDMODE('e', NULL, &opt,
 411                             N_("exit with zero when there's no error"), 'e'),
 412                 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print object's content"), 'p'),
 413                 OPT_CMDMODE(0, "textconv", &opt,
 414                             N_("for blob objects, run textconv on object's content"), 'c'),
 415                 OPT_BOOL( 0, "allow-unknown-type", &unknown_type,
 416                           N_("allow -s and -t to work with broken/corrupt objects")),
 417                 { OPTION_CALLBACK, 0, "batch", &batch, "format",
 418                         N_("show info and content of objects fed from the standard input"),
 419                         PARSE_OPT_OPTARG, batch_option_callback },
 420                 { OPTION_CALLBACK, 0, "batch-check", &batch, "format",
 421                         N_("show info about objects fed from the standard input"),
 422                         PARSE_OPT_OPTARG, batch_option_callback },
 423                 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
 424                          N_("follow in-tree symlinks (used with --batch or --batch-check)")),
 425                 OPT_END()
 426         };
 427
 428         git_config(git_cat_file_config, NULL);
 429
 430         argc = parse_options(argc, argv, prefix, options, cat_file_usage, 0);
 431
 432         if (opt) {
 433                 if (argc == 1)
 434                         obj_name = argv[0];
 435                 else
 436                         usage_with_options(cat_file_usage, options);
 437         }
 438         if (!opt && !batch.enabled) {
 439                 if (argc == 2) {
 440                         exp_type = argv[0];
 441                         obj_name = argv[1];
 442                 } else
 443                         usage_with_options(cat_file_usage, options);
 444         }
 445         if (batch.enabled && (opt || argc)) {
 446                 usage_with_options(cat_file_usage, options);
 447         }
 448
 449         if (batch.follow_symlinks && !batch.enabled) {
 450                 usage_with_options(cat_file_usage, options);
 451         }
 452
 453         if (batch.enabled)
 454                 return batch_objects(&batch);
 455
 456         if (unknown_type && opt != 't' && opt != 's')
 457                 die("git cat-file --allow-unknown-type: use with -s or -t");
 458         return cat_one_file(opt, exp_type, obj_name, unknown_type);
 459 }