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