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 struct batch_options {
  14         int enabled;
  15         int follow_symlinks;
  16         int print_contents;
  17         int buffer_output;
  18         int all_objects;
  19         const char *format;
  20 };
  21
  22 static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
  23                         int unknown_type)
  24 {
  25         unsigned char sha1[20];
  26         enum object_type type;
  27         char *buf;
  28         unsigned long size;
  29         struct object_context obj_context;
  30         struct object_info oi = {NULL};
  31         struct strbuf sb = STRBUF_INIT;
  32         unsigned flags = LOOKUP_REPLACE_OBJECT;
  33
  34         if (unknown_type)
  35                 flags |= LOOKUP_UNKNOWN_OBJECT;
  36
  37         if (get_sha1_with_context(obj_name, 0, sha1, &obj_context))
  38                 die("Not a valid object name %s", obj_name);
  39
  40         buf = NULL;
  41         switch (opt) {
  42         case 't':
  43                 oi.typename = &sb;
  44                 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
  45                         die("git cat-file: could not get object info");
  46                 if (sb.len) {
  47                         printf("%s\n", sb.buf);
  48                         strbuf_release(&sb);
  49                         return 0;
  50                 }
  51                 break;
  52
  53         case 's':
  54                 oi.sizep = &size;
  55                 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
  56                         die("git cat-file: could not get object info");
  57                 printf("%lu\n", size);
  58                 return 0;
  59
  60         case 'e':
  61                 return !has_sha1_file(sha1);
  62
  63         case 'c':
  64                 if (!obj_context.path[0])
  65                         die("git cat-file --textconv %s: <object> must be <sha1:path>",
  66                             obj_name);
  67
  68                 if (textconv_object(obj_context.path, obj_context.mode, sha1, 1, &buf, &size))
  69                         break;
  70
  71         case 'p':
  72                 type = sha1_object_info(sha1, NULL);
  73                 if (type < 0)
  74                         die("Not a valid object name %s", obj_name);
  75
  76                 /* custom pretty-print here */
  77                 if (type == OBJ_TREE) {
  78                         const char *ls_args[3] = { NULL };
  79                         ls_args[0] =  "ls-tree";
  80                         ls_args[1] =  obj_name;
  81                         return cmd_ls_tree(2, ls_args, NULL);
  82                 }
  83
  84                 if (type == OBJ_BLOB)
  85                         return stream_blob_to_fd(1, sha1, NULL, 0);
  86                 buf = read_sha1_file(sha1, &type, &size);
  87                 if (!buf)
  88                         die("Cannot read object %s", obj_name);
  89
  90                 /* otherwise just spit out the data */
  91                 break;
  92
  93         case 0:
  94                 if (type_from_string(exp_type) == OBJ_BLOB) {
  95                         unsigned char blob_sha1[20];
  96                         if (sha1_object_info(sha1, NULL) == OBJ_TAG) {
  97                                 char *buffer = read_sha1_file(sha1, &type, &size);
  98                                 const char *target;
  99                                 if (!skip_prefix(buffer, "object ", &target) ||
 100                                     get_sha1_hex(target, blob_sha1))
 101                                         die("%s not a valid tag", sha1_to_hex(sha1));
 102                                 free(buffer);
 103                         } else
 104                                 hashcpy(blob_sha1, sha1);
 105
 106                         if (sha1_object_info(blob_sha1, NULL) == OBJ_BLOB)
 107                                 return stream_blob_to_fd(1, blob_sha1, NULL, 0);
 108                         /*
 109                          * we attempted to dereference a tag to a blob
 110                          * and failed; there may be new dereference
 111                          * mechanisms this code is not aware of.
 112                          * fall-back to the usual case.
 113                          */
 114                 }
 115                 buf = read_object_with_reference(sha1, exp_type, &size, NULL);
 116                 break;
 117
 118         default:
 119                 die("git cat-file: unknown option: %s", exp_type);
 120         }
 121
 122         if (!buf)
 123                 die("git cat-file %s: bad file", obj_name);
 124
 125         write_or_die(1, buf, size);
 126         return 0;
 127 }
 128
 129 struct expand_data {
 130         unsigned char sha1[20];
 131         enum object_type type;
 132         unsigned long size;
 133         unsigned long disk_size;
 134         const char *rest;
 135         unsigned char delta_base_sha1[20];
 136
 137         /*
 138          * If mark_query is true, we do not expand anything, but rather
 139          * just mark the object_info with items we wish to query.
 140          */
 141         int mark_query;
 142
 143         /*
 144          * Whether to split the input on whitespace before feeding it to
 145          * get_sha1; this is decided during the mark_query phase based on
 146          * whether we have a %(rest) token in our format.
 147          */
 148         int split_on_whitespace;
 149
 150         /*
 151          * After a mark_query run, this object_info is set up to be
 152          * passed to sha1_object_info_extended. It will point to the data
 153          * elements above, so you can retrieve the response from there.
 154          */
 155         struct object_info info;
 156 };
 157
 158 static int is_atom(const char *atom, const char *s, int slen)
 159 {
 160         int alen = strlen(atom);
 161         return alen == slen && !memcmp(atom, s, alen);
 162 }
 163
 164 static void expand_atom(struct strbuf *sb, const char *atom, int len,
 165                         void *vdata)
 166 {
 167         struct expand_data *data = vdata;
 168
 169         if (is_atom("objectname", atom, len)) {
 170                 if (!data->mark_query)
 171                         strbuf_addstr(sb, sha1_to_hex(data->sha1));
 172         } else if (is_atom("objecttype", atom, len)) {
 173                 if (data->mark_query)
 174                         data->info.typep = &data->type;
 175                 else
 176                         strbuf_addstr(sb, typename(data->type));
 177         } else if (is_atom("objectsize", atom, len)) {
 178                 if (data->mark_query)
 179                         data->info.sizep = &data->size;
 180                 else
 181                         strbuf_addf(sb, "%lu", data->size);
 182         } else if (is_atom("objectsize:disk", atom, len)) {
 183                 if (data->mark_query)
 184                         data->info.disk_sizep = &data->disk_size;
 185                 else
 186                         strbuf_addf(sb, "%lu", data->disk_size);
 187         } else if (is_atom("rest", atom, len)) {
 188                 if (data->mark_query)
 189                         data->split_on_whitespace = 1;
 190                 else if (data->rest)
 191                         strbuf_addstr(sb, data->rest);
 192         } else if (is_atom("deltabase", atom, len)) {
 193                 if (data->mark_query)
 194                         data->info.delta_base_sha1 = data->delta_base_sha1;
 195                 else
 196                         strbuf_addstr(sb, sha1_to_hex(data->delta_base_sha1));
 197         } else
 198                 die("unknown format element: %.*s", len, atom);
 199 }
 200
 201 static size_t expand_format(struct strbuf *sb, const char *start, void *data)
 202 {
 203         const char *end;
 204
 205         if (*start != '(')
 206                 return 0;
 207         end = strchr(start + 1, ')');
 208         if (!end)
 209                 die("format element '%s' does not end in ')'", start);
 210
 211         expand_atom(sb, start + 1, end - start - 1, data);
 212
 213         return end - start + 1;
 214 }
 215
 216 static void batch_write(struct batch_options *opt, const void *data, int len)
 217 {
 218         if (opt->buffer_output) {
 219                 if (fwrite(data, 1, len, stdout) != len)
 220                         die_errno("unable to write to stdout");
 221         } else
 222                 write_or_die(1, data, len);
 223 }
 224
 225 static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
 226 {
 227         const unsigned char *sha1 = data->sha1;
 228
 229         assert(data->info.typep);
 230
 231         if (data->type == OBJ_BLOB) {
 232                 if (opt->buffer_output)
 233                         fflush(stdout);
 234                 if (stream_blob_to_fd(1, sha1, NULL, 0) < 0)
 235                         die("unable to stream %s to stdout", sha1_to_hex(sha1));
 236         }
 237         else {
 238                 enum object_type type;
 239                 unsigned long size;
 240                 void *contents;
 241
 242                 contents = read_sha1_file(sha1, &type, &size);
 243                 if (!contents)
 244                         die("object %s disappeared", sha1_to_hex(sha1));
 245                 if (type != data->type)
 246                         die("object %s changed type!?", sha1_to_hex(sha1));
 247                 if (data->info.sizep && size != data->size)
 248                         die("object %s changed size!?", sha1_to_hex(sha1));
 249
 250                 batch_write(opt, contents, size);
 251                 free(contents);
 252         }
 253 }
 254
 255 static void batch_object_write(const char *obj_name, struct batch_options *opt,
 256                                struct expand_data *data)
 257 {
 258         struct strbuf buf = STRBUF_INIT;
 259
 260         if (sha1_object_info_extended(data->sha1, &data->info, LOOKUP_REPLACE_OBJECT) < 0) {
 261                 printf("%s missing\n", obj_name ? obj_name : sha1_to_hex(data->sha1));
 262                 fflush(stdout);
 263                 return;
 264         }
 265
 266         strbuf_expand(&buf, opt->format, expand_format, data);
 267         strbuf_addch(&buf, '\n');
 268         batch_write(opt, buf.buf, buf.len);
 269         strbuf_release(&buf);
 270
 271         if (opt->print_contents) {
 272                 print_object_or_die(opt, data);
 273                 batch_write(opt, "\n", 1);
 274         }
 275 }
 276
 277 static void batch_one_object(const char *obj_name, struct batch_options *opt,
 278                              struct expand_data *data)
 279 {
 280         struct object_context ctx;
 281         int flags = opt->follow_symlinks ? GET_SHA1_FOLLOW_SYMLINKS : 0;
 282         enum follow_symlinks_result result;
 283
 284         result = get_sha1_with_context(obj_name, flags, data->sha1, &ctx);
 285         if (result != FOUND) {
 286                 switch (result) {
 287                 case MISSING_OBJECT:
 288                         printf("%s missing\n", obj_name);
 289                         break;
 290                 case DANGLING_SYMLINK:
 291                         printf("dangling %"PRIuMAX"\n%s\n",
 292                                (uintmax_t)strlen(obj_name), obj_name);
 293                         break;
 294                 case SYMLINK_LOOP:
 295                         printf("loop %"PRIuMAX"\n%s\n",
 296                                (uintmax_t)strlen(obj_name), obj_name);
 297                         break;
 298                 case NOT_DIR:
 299                         printf("notdir %"PRIuMAX"\n%s\n",
 300                                (uintmax_t)strlen(obj_name), obj_name);
 301                         break;
 302                 default:
 303                         die("BUG: unknown get_sha1_with_context result %d\n",
 304                                result);
 305                         break;
 306                 }
 307                 fflush(stdout);
 308                 return;
 309         }
 310
 311         if (ctx.mode == 0) {
 312                 printf("symlink %"PRIuMAX"\n%s\n",
 313                        (uintmax_t)ctx.symlink_path.len,
 314                        ctx.symlink_path.buf);
 315                 fflush(stdout);
 316                 return;
 317         }
 318
 319         batch_object_write(obj_name, opt, data);
 320 }
 321
 322 struct object_cb_data {
 323         struct batch_options *opt;
 324         struct expand_data *expand;
 325 };
 326
 327 static int batch_object_cb(const unsigned char *sha1,
 328                            struct object_cb_data *data)
 329 {
 330         hashcpy(data->expand->sha1, sha1);
 331         batch_object_write(NULL, data->opt, data->expand);
 332         return 0;
 333 }
 334
 335 static int batch_loose_object(const unsigned char *sha1,
 336                               const char *path,
 337                               void *data)
 338 {
 339         return batch_object_cb(sha1, data);
 340 }
 341
 342 static int batch_packed_object(const unsigned char *sha1,
 343                                struct packed_git *pack,
 344                                uint32_t pos,
 345                                void *data)
 346 {
 347         return batch_object_cb(sha1, data);
 348 }
 349
 350 static int batch_objects(struct batch_options *opt)
 351 {
 352         struct strbuf buf = STRBUF_INIT;
 353         struct expand_data data;
 354         int save_warning;
 355         int retval = 0;
 356
 357         if (!opt->format)
 358                 opt->format = "%(objectname) %(objecttype) %(objectsize)";
 359
 360         /*
 361          * Expand once with our special mark_query flag, which will prime the
 362          * object_info to be handed to sha1_object_info_extended for each
 363          * object.
 364          */
 365         memset(&data, 0, sizeof(data));
 366         data.mark_query = 1;
 367         strbuf_expand(&buf, opt->format, expand_format, &data);
 368         data.mark_query = 0;
 369
 370         /*
 371          * If we are printing out the object, then always fill in the type,
 372          * since we will want to decide whether or not to stream.
 373          */
 374         if (opt->print_contents)
 375                 data.info.typep = &data.type;
 376
 377         if (opt->all_objects) {
 378                 struct object_cb_data cb;
 379                 cb.opt = opt;
 380                 cb.expand = &data;
 381                 for_each_loose_object(batch_loose_object, &cb, 0);
 382                 for_each_packed_object(batch_packed_object, &cb, 0);
 383                 return 0;
 384         }
 385
 386         /*
 387          * We are going to call get_sha1 on a potentially very large number of
 388          * objects. In most large cases, these will be actual object sha1s. The
 389          * cost to double-check that each one is not also a ref (just so we can
 390          * warn) ends up dwarfing the actual cost of the object lookups
 391          * themselves. We can work around it by just turning off the warning.
 392          */
 393         save_warning = warn_on_object_refname_ambiguity;
 394         warn_on_object_refname_ambiguity = 0;
 395
 396         while (strbuf_getline(&buf, stdin, '\n') != EOF) {
 397                 if (data.split_on_whitespace) {
 398                         /*
 399                          * Split at first whitespace, tying off the beginning
 400                          * of the string and saving the remainder (or NULL) in
 401                          * data.rest.
 402                          */
 403                         char *p = strpbrk(buf.buf, " \t");
 404                         if (p) {
 405                                 while (*p && strchr(" \t", *p))
 406                                         *p++ = '\0';
 407                         }
 408                         data.rest = p;
 409                 }
 410
 411                 batch_one_object(buf.buf, opt, &data);
 412         }
 413
 414         strbuf_release(&buf);
 415         warn_on_object_refname_ambiguity = save_warning;
 416         return retval;
 417 }
 418
 419 static const char * const cat_file_usage[] = {
 420         N_("git cat-file (-t [--allow-unknown-type]|-s [--allow-unknown-type]|-e|-p|<type>|--textconv) <object>"),
 421         N_("git cat-file (--batch | --batch-check) [--follow-symlinks] < <list-of-objects>"),
 422         NULL
 423 };
 424
 425 static int git_cat_file_config(const char *var, const char *value, void *cb)
 426 {
 427         if (userdiff_config(var, value) < 0)
 428                 return -1;
 429
 430         return git_default_config(var, value, cb);
 431 }
 432
 433 static int batch_option_callback(const struct option *opt,
 434                                  const char *arg,
 435                                  int unset)
 436 {
 437         struct batch_options *bo = opt->value;
 438
 439         if (bo->enabled) {
 440                 return 1;
 441         }
 442
 443         bo->enabled = 1;
 444         bo->print_contents = !strcmp(opt->long_name, "batch");
 445         bo->format = arg;
 446
 447         return 0;
 448 }
 449
 450 int cmd_cat_file(int argc, const char **argv, const char *prefix)
 451 {
 452         int opt = 0;
 453         const char *exp_type = NULL, *obj_name = NULL;
 454         struct batch_options batch = {0};
 455         int unknown_type = 0;
 456
 457         const struct option options[] = {
 458                 OPT_GROUP(N_("<type> can be one of: blob, tree, commit, tag")),
 459                 OPT_CMDMODE('t', NULL, &opt, N_("show object type"), 't'),
 460                 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
 461                 OPT_CMDMODE('e', NULL, &opt,
 462                             N_("exit with zero when there's no error"), 'e'),
 463                 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print object's content"), 'p'),
 464                 OPT_CMDMODE(0, "textconv", &opt,
 465                             N_("for blob objects, run textconv on object's content"), 'c'),
 466                 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
 467                           N_("allow -s and -t to work with broken/corrupt objects")),
 468                 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
 469                 { OPTION_CALLBACK, 0, "batch", &batch, "format",
 470                         N_("show info and content of objects fed from the standard input"),
 471                         PARSE_OPT_OPTARG, batch_option_callback },
 472                 { OPTION_CALLBACK, 0, "batch-check", &batch, "format",
 473                         N_("show info about objects fed from the standard input"),
 474                         PARSE_OPT_OPTARG, batch_option_callback },
 475                 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
 476                          N_("follow in-tree symlinks (used with --batch or --batch-check)")),
 477                 OPT_BOOL(0, "batch-all-objects", &batch.all_objects,
 478                          N_("show all objects with --batch or --batch-check")),
 479                 OPT_END()
 480         };
 481
 482         git_config(git_cat_file_config, NULL);
 483
 484         argc = parse_options(argc, argv, prefix, options, cat_file_usage, 0);
 485
 486         if (opt) {
 487                 if (argc == 1)
 488                         obj_name = argv[0];
 489                 else
 490                         usage_with_options(cat_file_usage, options);
 491         }
 492         if (!opt && !batch.enabled) {
 493                 if (argc == 2) {
 494                         exp_type = argv[0];
 495                         obj_name = argv[1];
 496                 } else
 497                         usage_with_options(cat_file_usage, options);
 498         }
 499         if (batch.enabled && (opt || argc)) {
 500                 usage_with_options(cat_file_usage, options);
 501         }
 502
 503         if ((batch.follow_symlinks || batch.all_objects) && !batch.enabled) {
 504                 usage_with_options(cat_file_usage, options);
 505         }
 506
 507         if (batch.enabled)
 508                 return batch_objects(&batch);
 509
 510         if (unknown_type && opt != 't' && opt != 's')
 511                 die("git cat-file --allow-unknown-type: use with -s or -t");
 512         return cat_one_file(opt, exp_type, obj_name, unknown_type);
 513 }