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