cat-file: introduce the --filters option
[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"
3115ee45 12#include "sha1-array.h"
15d8e565 13
bfd15594
JK
14struct batch_options {
15 int enabled;
16 int follow_symlinks;
17 int print_contents;
fc4937c3 18 int buffer_output;
6a951937 19 int all_objects;
bfd15594
JK
20 const char *format;
21};
22
b9e62f60
JS
23static int filter_object(const char *path, unsigned mode,
24 const unsigned char *sha1,
25 char **buf, unsigned long *size)
26{
27 enum object_type type;
28
29 *buf = read_sha1_file(sha1, &type, size);
30 if (!*buf)
31 return error(_("cannot read object %s '%s'"),
32 sha1_to_hex(sha1), path);
33 if ((type == OBJ_BLOB) && S_ISREG(mode)) {
34 struct strbuf strbuf = STRBUF_INIT;
35 if (convert_to_working_tree(path, *buf, *size, &strbuf)) {
36 free(*buf);
37 *size = strbuf.len;
38 *buf = strbuf_detach(&strbuf, NULL);
39 }
40 }
41
42 return 0;
43}
44
39e4ae38
KN
45static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
46 int unknown_type)
e83c5163
LT
47{
48 unsigned char sha1[20];
21666f1a 49 enum object_type type;
e5fba602 50 char *buf;
e83c5163 51 unsigned long size;
e5fba602 52 struct object_context obj_context;
39e4ae38
KN
53 struct object_info oi = {NULL};
54 struct strbuf sb = STRBUF_INIT;
55 unsigned flags = LOOKUP_REPLACE_OBJECT;
56
57 if (unknown_type)
58 flags |= LOOKUP_UNKNOWN_OBJECT;
2b6854c8 59
33bd598c 60 if (get_sha1_with_context(obj_name, 0, sha1, &obj_context))
2b6854c8 61 die("Not a valid object name %s", obj_name);
7950571a 62
7950571a
PA
63 buf = NULL;
64 switch (opt) {
65 case 't':
39e4ae38
KN
66 oi.typename = &sb;
67 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
68 die("git cat-file: could not get object info");
69 if (sb.len) {
70 printf("%s\n", sb.buf);
71 strbuf_release(&sb);
f2a06330 72 return 0;
11e7d5c5 73 }
7950571a
PA
74 break;
75
76 case 's':
39e4ae38
KN
77 oi.sizep = &size;
78 if (sha1_object_info_extended(sha1, &oi, flags) < 0)
79 die("git cat-file: could not get object info");
80 printf("%lu\n", size);
81 return 0;
7950571a
PA
82
83 case 'e':
84 return !has_sha1_file(sha1);
85
b9e62f60
JS
86 case 'w':
87 if (!obj_context.path[0])
88 die("git cat-file --filters %s: <object> must be "
89 "<sha1:path>", obj_name);
90
91 if (filter_object(obj_context.path, obj_context.mode,
92 sha1, &buf, &size))
93 return -1;
94 break;
95
3ac21617
MG
96 case 'c':
97 if (!obj_context.path[0])
98 die("git cat-file --textconv %s: <object> must be <sha1:path>",
99 obj_name);
100
101 if (textconv_object(obj_context.path, obj_context.mode, sha1, 1, &buf, &size))
102 break;
103
a0f15fa5 104 case 'p':
21666f1a
NP
105 type = sha1_object_info(sha1, NULL);
106 if (type < 0)
2b6854c8 107 die("Not a valid object name %s", obj_name);
a0f15fa5
JH
108
109 /* custom pretty-print here */
2b6854c8 110 if (type == OBJ_TREE) {
66dbfd55
GV
111 const char *ls_args[3] = { NULL };
112 ls_args[0] = "ls-tree";
113 ls_args[1] = obj_name;
2b6854c8
SP
114 return cmd_ls_tree(2, ls_args, NULL);
115 }
a0f15fa5 116
00c8fd49
NTND
117 if (type == OBJ_BLOB)
118 return stream_blob_to_fd(1, sha1, NULL, 0);
21666f1a 119 buf = read_sha1_file(sha1, &type, &size);
a0f15fa5 120 if (!buf)
2b6854c8 121 die("Cannot read object %s", obj_name);
a0f15fa5
JH
122
123 /* otherwise just spit out the data */
124 break;
e5fba602 125
7950571a 126 case 0:
00c8fd49
NTND
127 if (type_from_string(exp_type) == OBJ_BLOB) {
128 unsigned char blob_sha1[20];
129 if (sha1_object_info(sha1, NULL) == OBJ_TAG) {
00c8fd49 130 char *buffer = read_sha1_file(sha1, &type, &size);
e3f1da98
RS
131 const char *target;
132 if (!skip_prefix(buffer, "object ", &target) ||
133 get_sha1_hex(target, blob_sha1))
00c8fd49
NTND
134 die("%s not a valid tag", sha1_to_hex(sha1));
135 free(buffer);
136 } else
137 hashcpy(blob_sha1, sha1);
138
139 if (sha1_object_info(blob_sha1, NULL) == OBJ_BLOB)
140 return stream_blob_to_fd(1, blob_sha1, NULL, 0);
141 /*
142 * we attempted to dereference a tag to a blob
143 * and failed; there may be new dereference
144 * mechanisms this code is not aware of.
145 * fall-back to the usual case.
146 */
147 }
2b6854c8 148 buf = read_object_with_reference(sha1, exp_type, &size, NULL);
7950571a
PA
149 break;
150
151 default:
d7530708 152 die("git cat-file: unknown option: %s", exp_type);
bf0c6e83
LT
153 }
154
11e7d5c5 155 if (!buf)
34baebce 156 die("git cat-file %s: bad file", obj_name);
11e7d5c5 157
7230e6d0 158 write_or_die(1, buf, size);
bf0c6e83 159 return 0;
e83c5163 160}
9cf71b17 161
93d2a607
JK
162struct expand_data {
163 unsigned char sha1[20];
164 enum object_type type;
165 unsigned long size;
166df26f 166 off_t disk_size;
97be0407 167 const char *rest;
65ea9c3c 168 unsigned char delta_base_sha1[20];
93d2a607
JK
169
170 /*
171 * If mark_query is true, we do not expand anything, but rather
172 * just mark the object_info with items we wish to query.
173 */
174 int mark_query;
175
97be0407
JK
176 /*
177 * Whether to split the input on whitespace before feeding it to
178 * get_sha1; this is decided during the mark_query phase based on
179 * whether we have a %(rest) token in our format.
180 */
181 int split_on_whitespace;
182
93d2a607
JK
183 /*
184 * After a mark_query run, this object_info is set up to be
185 * passed to sha1_object_info_extended. It will point to the data
186 * elements above, so you can retrieve the response from there.
187 */
188 struct object_info info;
845de33a
JK
189
190 /*
191 * This flag will be true if the requested batch format and options
192 * don't require us to call sha1_object_info, which can then be
193 * optimized out.
194 */
195 unsigned skip_object_info : 1;
93d2a607
JK
196};
197
198static int is_atom(const char *atom, const char *s, int slen)
199{
200 int alen = strlen(atom);
201 return alen == slen && !memcmp(atom, s, alen);
202}
203
204static void expand_atom(struct strbuf *sb, const char *atom, int len,
205 void *vdata)
206{
207 struct expand_data *data = vdata;
208
209 if (is_atom("objectname", atom, len)) {
210 if (!data->mark_query)
211 strbuf_addstr(sb, sha1_to_hex(data->sha1));
212 } else if (is_atom("objecttype", atom, len)) {
5b086407
JK
213 if (data->mark_query)
214 data->info.typep = &data->type;
215 else
93d2a607
JK
216 strbuf_addstr(sb, typename(data->type));
217 } else if (is_atom("objectsize", atom, len)) {
218 if (data->mark_query)
219 data->info.sizep = &data->size;
220 else
221 strbuf_addf(sb, "%lu", data->size);
a4ac1061
JK
222 } else if (is_atom("objectsize:disk", atom, len)) {
223 if (data->mark_query)
224 data->info.disk_sizep = &data->disk_size;
225 else
166df26f 226 strbuf_addf(sb, "%"PRIuMAX, (uintmax_t)data->disk_size);
97be0407
JK
227 } else if (is_atom("rest", atom, len)) {
228 if (data->mark_query)
229 data->split_on_whitespace = 1;
230 else if (data->rest)
231 strbuf_addstr(sb, data->rest);
65ea9c3c
JK
232 } else if (is_atom("deltabase", atom, len)) {
233 if (data->mark_query)
234 data->info.delta_base_sha1 = data->delta_base_sha1;
235 else
236 strbuf_addstr(sb, sha1_to_hex(data->delta_base_sha1));
93d2a607
JK
237 } else
238 die("unknown format element: %.*s", len, atom);
239}
240
241static size_t expand_format(struct strbuf *sb, const char *start, void *data)
242{
243 const char *end;
244
245 if (*start != '(')
246 return 0;
247 end = strchr(start + 1, ')');
248 if (!end)
249 die("format element '%s' does not end in ')'", start);
250
251 expand_atom(sb, start + 1, end - start - 1, data);
252
253 return end - start + 1;
254}
255
fc4937c3
JK
256static void batch_write(struct batch_options *opt, const void *data, int len)
257{
258 if (opt->buffer_output) {
259 if (fwrite(data, 1, len, stdout) != len)
260 die_errno("unable to write to stdout");
261 } else
262 write_or_die(1, data, len);
263}
264
265static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
98e2092b 266{
370c9268
JK
267 const unsigned char *sha1 = data->sha1;
268
6554dfa9
JK
269 assert(data->info.typep);
270
370c9268 271 if (data->type == OBJ_BLOB) {
fc4937c3
JK
272 if (opt->buffer_output)
273 fflush(stdout);
274 if (stream_blob_to_fd(1, sha1, NULL, 0) < 0)
98e2092b
JK
275 die("unable to stream %s to stdout", sha1_to_hex(sha1));
276 }
277 else {
370c9268
JK
278 enum object_type type;
279 unsigned long size;
98e2092b
JK
280 void *contents;
281
370c9268 282 contents = read_sha1_file(sha1, &type, &size);
98e2092b
JK
283 if (!contents)
284 die("object %s disappeared", sha1_to_hex(sha1));
370c9268 285 if (type != data->type)
98e2092b 286 die("object %s changed type!?", sha1_to_hex(sha1));
6554dfa9 287 if (data->info.sizep && size != data->size)
370c9268 288 die("object %s changed size!?", sha1_to_hex(sha1));
98e2092b 289
fc4937c3 290 batch_write(opt, contents, size);
98e2092b
JK
291 free(contents);
292 }
293}
294
44b877e9
JK
295static void batch_object_write(const char *obj_name, struct batch_options *opt,
296 struct expand_data *data)
297{
298 struct strbuf buf = STRBUF_INIT;
299
845de33a
JK
300 if (!data->skip_object_info &&
301 sha1_object_info_extended(data->sha1, &data->info, LOOKUP_REPLACE_OBJECT) < 0) {
6a951937 302 printf("%s missing\n", obj_name ? obj_name : sha1_to_hex(data->sha1));
44b877e9
JK
303 fflush(stdout);
304 return;
305 }
306
307 strbuf_expand(&buf, opt->format, expand_format, data);
308 strbuf_addch(&buf, '\n');
309 batch_write(opt, buf.buf, buf.len);
310 strbuf_release(&buf);
311
312 if (opt->print_contents) {
313 print_object_or_die(opt, data);
314 batch_write(opt, "\n", 1);
315 }
316}
317
82330950
JK
318static void batch_one_object(const char *obj_name, struct batch_options *opt,
319 struct expand_data *data)
05d5667f 320{
122d5346
DT
321 struct object_context ctx;
322 int flags = opt->follow_symlinks ? GET_SHA1_FOLLOW_SYMLINKS : 0;
323 enum follow_symlinks_result result;
05d5667f 324
122d5346
DT
325 result = get_sha1_with_context(obj_name, flags, data->sha1, &ctx);
326 if (result != FOUND) {
327 switch (result) {
328 case MISSING_OBJECT:
329 printf("%s missing\n", obj_name);
330 break;
331 case DANGLING_SYMLINK:
332 printf("dangling %"PRIuMAX"\n%s\n",
333 (uintmax_t)strlen(obj_name), obj_name);
334 break;
335 case SYMLINK_LOOP:
336 printf("loop %"PRIuMAX"\n%s\n",
337 (uintmax_t)strlen(obj_name), obj_name);
338 break;
339 case NOT_DIR:
340 printf("notdir %"PRIuMAX"\n%s\n",
341 (uintmax_t)strlen(obj_name), obj_name);
342 break;
343 default:
344 die("BUG: unknown get_sha1_with_context result %d\n",
345 result);
346 break;
347 }
348 fflush(stdout);
82330950 349 return;
122d5346
DT
350 }
351
352 if (ctx.mode == 0) {
353 printf("symlink %"PRIuMAX"\n%s\n",
354 (uintmax_t)ctx.symlink_path.len,
355 ctx.symlink_path.buf);
422b2063 356 fflush(stdout);
82330950 357 return;
05d5667f
AR
358 }
359
44b877e9 360 batch_object_write(obj_name, opt, data);
05d5667f
AR
361}
362
6a951937
JK
363struct object_cb_data {
364 struct batch_options *opt;
365 struct expand_data *expand;
366};
367
3115ee45 368static void batch_object_cb(const unsigned char sha1[20], void *vdata)
6a951937 369{
3115ee45 370 struct object_cb_data *data = vdata;
6a951937
JK
371 hashcpy(data->expand->sha1, sha1);
372 batch_object_write(NULL, data->opt, data->expand);
6a951937
JK
373}
374
375static int batch_loose_object(const unsigned char *sha1,
376 const char *path,
377 void *data)
378{
3115ee45
JK
379 sha1_array_append(data, sha1);
380 return 0;
6a951937
JK
381}
382
383static int batch_packed_object(const unsigned char *sha1,
384 struct packed_git *pack,
385 uint32_t pos,
386 void *data)
387{
3115ee45
JK
388 sha1_array_append(data, sha1);
389 return 0;
6a951937
JK
390}
391
b71bd480 392static int batch_objects(struct batch_options *opt)
05d5667f 393{
f285a2d7 394 struct strbuf buf = STRBUF_INIT;
93d2a607 395 struct expand_data data;
a42fcd15 396 int save_warning;
07e23839 397 int retval = 0;
93d2a607
JK
398
399 if (!opt->format)
400 opt->format = "%(objectname) %(objecttype) %(objectsize)";
401
402 /*
403 * Expand once with our special mark_query flag, which will prime the
404 * object_info to be handed to sha1_object_info_extended for each
405 * object.
406 */
407 memset(&data, 0, sizeof(data));
408 data.mark_query = 1;
409 strbuf_expand(&buf, opt->format, expand_format, &data);
410 data.mark_query = 0;
05d5667f 411
845de33a
JK
412 if (opt->all_objects) {
413 struct object_info empty;
414 memset(&empty, 0, sizeof(empty));
415 if (!memcmp(&data.info, &empty, sizeof(empty)))
416 data.skip_object_info = 1;
417 }
418
6554dfa9
JK
419 /*
420 * If we are printing out the object, then always fill in the type,
421 * since we will want to decide whether or not to stream.
422 */
423 if (opt->print_contents)
424 data.info.typep = &data.type;
425
6a951937 426 if (opt->all_objects) {
3115ee45 427 struct sha1_array sa = SHA1_ARRAY_INIT;
6a951937 428 struct object_cb_data cb;
3115ee45
JK
429
430 for_each_loose_object(batch_loose_object, &sa, 0);
431 for_each_packed_object(batch_packed_object, &sa, 0);
432
6a951937
JK
433 cb.opt = opt;
434 cb.expand = &data;
3115ee45
JK
435 sha1_array_for_each_unique(&sa, batch_object_cb, &cb);
436
437 sha1_array_clear(&sa);
6a951937
JK
438 return 0;
439 }
440
25fba78d
JK
441 /*
442 * We are going to call get_sha1 on a potentially very large number of
443 * objects. In most large cases, these will be actual object sha1s. The
444 * cost to double-check that each one is not also a ref (just so we can
445 * warn) ends up dwarfing the actual cost of the object lookups
446 * themselves. We can work around it by just turning off the warning.
447 */
a42fcd15 448 save_warning = warn_on_object_refname_ambiguity;
25fba78d
JK
449 warn_on_object_refname_ambiguity = 0;
450
b42ca3dd 451 while (strbuf_getline(&buf, stdin) != EOF) {
97be0407
JK
452 if (data.split_on_whitespace) {
453 /*
454 * Split at first whitespace, tying off the beginning
455 * of the string and saving the remainder (or NULL) in
456 * data.rest.
457 */
458 char *p = strpbrk(buf.buf, " \t");
459 if (p) {
460 while (*p && strchr(" \t", *p))
461 *p++ = '\0';
462 }
463 data.rest = p;
464 }
465
82330950 466 batch_one_object(buf.buf, opt, &data);
05d5667f
AR
467 }
468
648027c4 469 strbuf_release(&buf);
a42fcd15 470 warn_on_object_refname_ambiguity = save_warning;
07e23839 471 return retval;
05d5667f
AR
472}
473
15d8e565 474static const char * const cat_file_usage[] = {
b9e62f60 475 N_("git cat-file (-t [--allow-unknown-type]|-s [--allow-unknown-type]|-e|-p|<type>|--textconv|--filters) <object>"),
33e8fc87 476 N_("git cat-file (--batch | --batch-check) [--follow-symlinks]"),
15d8e565
MB
477 NULL
478};
4814dbe8 479
e5fba602
CP
480static int git_cat_file_config(const char *var, const char *value, void *cb)
481{
6680a087 482 if (userdiff_config(var, value) < 0)
e5fba602 483 return -1;
e5fba602
CP
484
485 return git_default_config(var, value, cb);
486}
487
b71bd480
JK
488static int batch_option_callback(const struct option *opt,
489 const char *arg,
490 int unset)
491{
492 struct batch_options *bo = opt->value;
493
122d5346
DT
494 if (bo->enabled) {
495 return 1;
b71bd480
JK
496 }
497
498 bo->enabled = 1;
499 bo->print_contents = !strcmp(opt->long_name, "batch");
93d2a607 500 bo->format = arg;
b71bd480
JK
501
502 return 0;
503}
504
9cf71b17
AR
505int cmd_cat_file(int argc, const char **argv, const char *prefix)
506{
b71bd480 507 int opt = 0;
4814dbe8 508 const char *exp_type = NULL, *obj_name = NULL;
b71bd480 509 struct batch_options batch = {0};
39e4ae38 510 int unknown_type = 0;
9cf71b17 511
15d8e565 512 const struct option options[] = {
d68faec7 513 OPT_GROUP(N_("<type> can be one of: blob, tree, commit, tag")),
b48158ac
KN
514 OPT_CMDMODE('t', NULL, &opt, N_("show object type"), 't'),
515 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
516 OPT_CMDMODE('e', NULL, &opt,
d68faec7 517 N_("exit with zero when there's no error"), 'e'),
b48158ac
KN
518 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print object's content"), 'p'),
519 OPT_CMDMODE(0, "textconv", &opt,
d68faec7 520 N_("for blob objects, run textconv on object's content"), 'c'),
b9e62f60
JS
521 OPT_CMDMODE(0, "filters", &opt,
522 N_("for blob objects, run filters on object's content"), 'w'),
ad42f28d 523 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
39e4ae38 524 N_("allow -s and -t to work with broken/corrupt objects")),
fc4937c3 525 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
93d2a607 526 { OPTION_CALLBACK, 0, "batch", &batch, "format",
b71bd480 527 N_("show info and content of objects fed from the standard input"),
93d2a607
JK
528 PARSE_OPT_OPTARG, batch_option_callback },
529 { OPTION_CALLBACK, 0, "batch-check", &batch, "format",
b71bd480 530 N_("show info about objects fed from the standard input"),
93d2a607 531 PARSE_OPT_OPTARG, batch_option_callback },
122d5346
DT
532 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
533 N_("follow in-tree symlinks (used with --batch or --batch-check)")),
6a951937
JK
534 OPT_BOOL(0, "batch-all-objects", &batch.all_objects,
535 N_("show all objects with --batch or --batch-check")),
15d8e565
MB
536 OPT_END()
537 };
a8128ed6 538
e5fba602 539 git_config(git_cat_file_config, NULL);
4814dbe8 540
6a36e1e7 541 batch.buffer_output = -1;
37782920 542 argc = parse_options(argc, argv, prefix, options, cat_file_usage, 0);
05d5667f 543
15d8e565
MB
544 if (opt) {
545 if (argc == 1)
546 obj_name = argv[0];
547 else
548 usage_with_options(cat_file_usage, options);
549 }
b71bd480 550 if (!opt && !batch.enabled) {
15d8e565
MB
551 if (argc == 2) {
552 exp_type = argv[0];
553 obj_name = argv[1];
554 } else
555 usage_with_options(cat_file_usage, options);
556 }
b71bd480 557 if (batch.enabled && (opt || argc)) {
15d8e565 558 usage_with_options(cat_file_usage, options);
9cf71b17
AR
559 }
560
6a951937 561 if ((batch.follow_symlinks || batch.all_objects) && !batch.enabled) {
122d5346
DT
562 usage_with_options(cat_file_usage, options);
563 }
564
6a36e1e7
JK
565 if (batch.buffer_output < 0)
566 batch.buffer_output = batch.all_objects;
567
b71bd480
JK
568 if (batch.enabled)
569 return batch_objects(&batch);
05d5667f 570
39e4ae38
KN
571 if (unknown_type && opt != 't' && opt != 's')
572 die("git cat-file --allow-unknown-type: use with -s or -t");
573 return cat_one_file(opt, exp_type, obj_name, unknown_type);
9cf71b17 574}