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