cat-file: split batch_one_object into two stages
[git/git.git] / builtin / cat-file.c
CommitLineData
8bc9a0c7
LT
1/*
2 * GIT - The information manager from hell
3 *
4 * Copyright (C) Linus Torvalds, 2005
5 */
e83c5163 6#include "cache.h"
f81daefe 7#include "builtin.h"
15d8e565 8#include "parse-options.h"
e5fba602 9#include "userdiff.h"
00c8fd49 10#include "streaming.h"
122d5346 11#include "tree-walk.h"
15d8e565 12
bfd15594
JK
13struct batch_options {
14 int enabled;
15 int follow_symlinks;
16 int print_contents;
fc4937c3 17 int buffer_output;
bfd15594
JK
18 const char *format;
19};
20
39e4ae38
KN
21static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
22 int unknown_type)
e83c5163
LT
23{
24 unsigned char sha1[20];
21666f1a 25 enum object_type type;
e5fba602 26 char *buf;
e83c5163 27 unsigned long size;
e5fba602 28 struct object_context obj_context;
39e4ae38
KN
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;
2b6854c8 35
33bd598c 36 if (get_sha1_with_context(obj_name, 0, sha1, &obj_context))
2b6854c8 37 die("Not a valid object name %s", obj_name);
7950571a 38
7950571a
PA
39 buf = NULL;
40 switch (opt) {
41 case 't':
39e4ae38
KN
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);
f2a06330 48 return 0;
11e7d5c5 49 }
7950571a
PA
50 break;
51
52 case 's':
39e4ae38
KN
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;
7950571a
PA
58
59 case 'e':
60 return !has_sha1_file(sha1);
61
3ac21617
MG
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
a0f15fa5 70 case 'p':
21666f1a
NP
71 type = sha1_object_info(sha1, NULL);
72 if (type < 0)
2b6854c8 73 die("Not a valid object name %s", obj_name);
a0f15fa5
JH
74
75 /* custom pretty-print here */
2b6854c8 76 if (type == OBJ_TREE) {
66dbfd55
GV
77 const char *ls_args[3] = { NULL };
78 ls_args[0] = "ls-tree";
79 ls_args[1] = obj_name;
2b6854c8
SP
80 return cmd_ls_tree(2, ls_args, NULL);
81 }
a0f15fa5 82
00c8fd49
NTND
83 if (type == OBJ_BLOB)
84 return stream_blob_to_fd(1, sha1, NULL, 0);
21666f1a 85 buf = read_sha1_file(sha1, &type, &size);
a0f15fa5 86 if (!buf)
2b6854c8 87 die("Cannot read object %s", obj_name);
a0f15fa5
JH
88
89 /* otherwise just spit out the data */
90 break;
e5fba602 91
7950571a 92 case 0:
00c8fd49
NTND
93 if (type_from_string(exp_type) == OBJ_BLOB) {
94 unsigned char blob_sha1[20];
95 if (sha1_object_info(sha1, NULL) == OBJ_TAG) {
00c8fd49 96 char *buffer = read_sha1_file(sha1, &type, &size);
e3f1da98
RS
97 const char *target;
98 if (!skip_prefix(buffer, "object ", &target) ||
99 get_sha1_hex(target, blob_sha1))
00c8fd49
NTND
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 }
2b6854c8 114 buf = read_object_with_reference(sha1, exp_type, &size, NULL);
7950571a
PA
115 break;
116
117 default:
d7530708 118 die("git cat-file: unknown option: %s", exp_type);
bf0c6e83
LT
119 }
120
11e7d5c5 121 if (!buf)
34baebce 122 die("git cat-file %s: bad file", obj_name);
11e7d5c5 123
7230e6d0 124 write_or_die(1, buf, size);
bf0c6e83 125 return 0;
e83c5163 126}
9cf71b17 127
93d2a607
JK
128struct expand_data {
129 unsigned char sha1[20];
130 enum object_type type;
131 unsigned long size;
a4ac1061 132 unsigned long disk_size;
97be0407 133 const char *rest;
65ea9c3c 134 unsigned char delta_base_sha1[20];
93d2a607
JK
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
97be0407
JK
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
93d2a607
JK
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
157static 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
163static 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)) {
5b086407
JK
172 if (data->mark_query)
173 data->info.typep = &data->type;
174 else
93d2a607
JK
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);
a4ac1061
JK
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);
97be0407
JK
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);
65ea9c3c
JK
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));
93d2a607
JK
196 } else
197 die("unknown format element: %.*s", len, atom);
198}
199
200static 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
fc4937c3
JK
215static 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
224static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
98e2092b 225{
370c9268
JK
226 const unsigned char *sha1 = data->sha1;
227
6554dfa9
JK
228 assert(data->info.typep);
229
370c9268 230 if (data->type == OBJ_BLOB) {
fc4937c3
JK
231 if (opt->buffer_output)
232 fflush(stdout);
233 if (stream_blob_to_fd(1, sha1, NULL, 0) < 0)
98e2092b
JK
234 die("unable to stream %s to stdout", sha1_to_hex(sha1));
235 }
236 else {
370c9268
JK
237 enum object_type type;
238 unsigned long size;
98e2092b
JK
239 void *contents;
240
370c9268 241 contents = read_sha1_file(sha1, &type, &size);
98e2092b
JK
242 if (!contents)
243 die("object %s disappeared", sha1_to_hex(sha1));
370c9268 244 if (type != data->type)
98e2092b 245 die("object %s changed type!?", sha1_to_hex(sha1));
6554dfa9 246 if (data->info.sizep && size != data->size)
370c9268 247 die("object %s changed size!?", sha1_to_hex(sha1));
98e2092b 248
fc4937c3 249 batch_write(opt, contents, size);
98e2092b
JK
250 free(contents);
251 }
252}
253
44b877e9
JK
254static void batch_object_write(const char *obj_name, struct batch_options *opt,
255 struct expand_data *data)
256{
257 struct strbuf buf = STRBUF_INIT;
258
259 if (sha1_object_info_extended(data->sha1, &data->info, LOOKUP_REPLACE_OBJECT) < 0) {
260 printf("%s missing\n", obj_name);
261 fflush(stdout);
262 return;
263 }
264
265 strbuf_expand(&buf, opt->format, expand_format, data);
266 strbuf_addch(&buf, '\n');
267 batch_write(opt, buf.buf, buf.len);
268 strbuf_release(&buf);
269
270 if (opt->print_contents) {
271 print_object_or_die(opt, data);
272 batch_write(opt, "\n", 1);
273 }
274}
275
82330950
JK
276static void batch_one_object(const char *obj_name, struct batch_options *opt,
277 struct expand_data *data)
05d5667f 278{
122d5346
DT
279 struct object_context ctx;
280 int flags = opt->follow_symlinks ? GET_SHA1_FOLLOW_SYMLINKS : 0;
281 enum follow_symlinks_result result;
05d5667f 282
122d5346
DT
283 result = get_sha1_with_context(obj_name, flags, data->sha1, &ctx);
284 if (result != FOUND) {
285 switch (result) {
286 case MISSING_OBJECT:
287 printf("%s missing\n", obj_name);
288 break;
289 case DANGLING_SYMLINK:
290 printf("dangling %"PRIuMAX"\n%s\n",
291 (uintmax_t)strlen(obj_name), obj_name);
292 break;
293 case SYMLINK_LOOP:
294 printf("loop %"PRIuMAX"\n%s\n",
295 (uintmax_t)strlen(obj_name), obj_name);
296 break;
297 case NOT_DIR:
298 printf("notdir %"PRIuMAX"\n%s\n",
299 (uintmax_t)strlen(obj_name), obj_name);
300 break;
301 default:
302 die("BUG: unknown get_sha1_with_context result %d\n",
303 result);
304 break;
305 }
306 fflush(stdout);
82330950 307 return;
122d5346
DT
308 }
309
310 if (ctx.mode == 0) {
311 printf("symlink %"PRIuMAX"\n%s\n",
312 (uintmax_t)ctx.symlink_path.len,
313 ctx.symlink_path.buf);
422b2063 314 fflush(stdout);
82330950 315 return;
05d5667f
AR
316 }
317
44b877e9 318 batch_object_write(obj_name, opt, data);
05d5667f
AR
319}
320
b71bd480 321static int batch_objects(struct batch_options *opt)
05d5667f 322{
f285a2d7 323 struct strbuf buf = STRBUF_INIT;
93d2a607 324 struct expand_data data;
a42fcd15 325 int save_warning;
07e23839 326 int retval = 0;
93d2a607
JK
327
328 if (!opt->format)
329 opt->format = "%(objectname) %(objecttype) %(objectsize)";
330
331 /*
332 * Expand once with our special mark_query flag, which will prime the
333 * object_info to be handed to sha1_object_info_extended for each
334 * object.
335 */
336 memset(&data, 0, sizeof(data));
337 data.mark_query = 1;
338 strbuf_expand(&buf, opt->format, expand_format, &data);
339 data.mark_query = 0;
05d5667f 340
6554dfa9
JK
341 /*
342 * If we are printing out the object, then always fill in the type,
343 * since we will want to decide whether or not to stream.
344 */
345 if (opt->print_contents)
346 data.info.typep = &data.type;
347
25fba78d
JK
348 /*
349 * We are going to call get_sha1 on a potentially very large number of
350 * objects. In most large cases, these will be actual object sha1s. The
351 * cost to double-check that each one is not also a ref (just so we can
352 * warn) ends up dwarfing the actual cost of the object lookups
353 * themselves. We can work around it by just turning off the warning.
354 */
a42fcd15 355 save_warning = warn_on_object_refname_ambiguity;
25fba78d
JK
356 warn_on_object_refname_ambiguity = 0;
357
05d5667f 358 while (strbuf_getline(&buf, stdin, '\n') != EOF) {
97be0407
JK
359 if (data.split_on_whitespace) {
360 /*
361 * Split at first whitespace, tying off the beginning
362 * of the string and saving the remainder (or NULL) in
363 * data.rest.
364 */
365 char *p = strpbrk(buf.buf, " \t");
366 if (p) {
367 while (*p && strchr(" \t", *p))
368 *p++ = '\0';
369 }
370 data.rest = p;
371 }
372
82330950 373 batch_one_object(buf.buf, opt, &data);
05d5667f
AR
374 }
375
648027c4 376 strbuf_release(&buf);
a42fcd15 377 warn_on_object_refname_ambiguity = save_warning;
07e23839 378 return retval;
05d5667f
AR
379}
380
15d8e565 381static const char * const cat_file_usage[] = {
39e4ae38 382 N_("git cat-file (-t [--allow-unknown-type]|-s [--allow-unknown-type]|-e|-p|<type>|--textconv) <object>"),
122d5346 383 N_("git cat-file (--batch | --batch-check) [--follow-symlinks] < <list-of-objects>"),
15d8e565
MB
384 NULL
385};
4814dbe8 386
e5fba602
CP
387static int git_cat_file_config(const char *var, const char *value, void *cb)
388{
6680a087 389 if (userdiff_config(var, value) < 0)
e5fba602 390 return -1;
e5fba602
CP
391
392 return git_default_config(var, value, cb);
393}
394
b71bd480
JK
395static int batch_option_callback(const struct option *opt,
396 const char *arg,
397 int unset)
398{
399 struct batch_options *bo = opt->value;
400
122d5346
DT
401 if (bo->enabled) {
402 return 1;
b71bd480
JK
403 }
404
405 bo->enabled = 1;
406 bo->print_contents = !strcmp(opt->long_name, "batch");
93d2a607 407 bo->format = arg;
b71bd480
JK
408
409 return 0;
410}
411
9cf71b17
AR
412int cmd_cat_file(int argc, const char **argv, const char *prefix)
413{
b71bd480 414 int opt = 0;
4814dbe8 415 const char *exp_type = NULL, *obj_name = NULL;
b71bd480 416 struct batch_options batch = {0};
39e4ae38 417 int unknown_type = 0;
9cf71b17 418
15d8e565 419 const struct option options[] = {
d68faec7 420 OPT_GROUP(N_("<type> can be one of: blob, tree, commit, tag")),
b48158ac
KN
421 OPT_CMDMODE('t', NULL, &opt, N_("show object type"), 't'),
422 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
423 OPT_CMDMODE('e', NULL, &opt,
d68faec7 424 N_("exit with zero when there's no error"), 'e'),
b48158ac
KN
425 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print object's content"), 'p'),
426 OPT_CMDMODE(0, "textconv", &opt,
d68faec7 427 N_("for blob objects, run textconv on object's content"), 'c'),
ad42f28d 428 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
39e4ae38 429 N_("allow -s and -t to work with broken/corrupt objects")),
fc4937c3 430 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
93d2a607 431 { OPTION_CALLBACK, 0, "batch", &batch, "format",
b71bd480 432 N_("show info and content of objects fed from the standard input"),
93d2a607
JK
433 PARSE_OPT_OPTARG, batch_option_callback },
434 { OPTION_CALLBACK, 0, "batch-check", &batch, "format",
b71bd480 435 N_("show info about objects fed from the standard input"),
93d2a607 436 PARSE_OPT_OPTARG, batch_option_callback },
122d5346
DT
437 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
438 N_("follow in-tree symlinks (used with --batch or --batch-check)")),
15d8e565
MB
439 OPT_END()
440 };
a8128ed6 441
e5fba602 442 git_config(git_cat_file_config, NULL);
4814dbe8 443
37782920 444 argc = parse_options(argc, argv, prefix, options, cat_file_usage, 0);
05d5667f 445
15d8e565
MB
446 if (opt) {
447 if (argc == 1)
448 obj_name = argv[0];
449 else
450 usage_with_options(cat_file_usage, options);
451 }
b71bd480 452 if (!opt && !batch.enabled) {
15d8e565
MB
453 if (argc == 2) {
454 exp_type = argv[0];
455 obj_name = argv[1];
456 } else
457 usage_with_options(cat_file_usage, options);
458 }
b71bd480 459 if (batch.enabled && (opt || argc)) {
15d8e565 460 usage_with_options(cat_file_usage, options);
9cf71b17
AR
461 }
462
122d5346
DT
463 if (batch.follow_symlinks && !batch.enabled) {
464 usage_with_options(cat_file_usage, options);
465 }
466
b71bd480
JK
467 if (batch.enabled)
468 return batch_objects(&batch);
05d5667f 469
39e4ae38
KN
470 if (unknown_type && opt != 't' && opt != 's')
471 die("git cat-file --allow-unknown-type: use with -s or -t");
472 return cat_one_file(opt, exp_type, obj_name, unknown_type);
9cf71b17 473}