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