?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
pkt-line: re-'static'-ify buffer in packet_write_fmt_1()
[git.git] / convert.c
blobbc242276ff095845f5a51d7d6ec3a9e97f02ddd0
1 #include "cache.h"
2 #include "attr.h"
3 #include "run-command.h"
4 #include "quote.h"
5 #include "sigchain.h"
6 #include "pkt-line.h"
7
8 /*
9 * convert.c - convert a file when checking it out and checking it in.
10 *
11 * This should use the pathname to decide on whether it wants to do some
12 * more interesting conversions (automatic gzip/unzip, general format
13 * conversions etc etc), but by default it just does automatic CRLF<->LF
14 * translation when the "text" attribute or "auto_crlf" option is set.
15 */
16
17 /* Stat bits: When BIN is set, the txt bits are unset */
18 #define CONVERT_STAT_BITS_TXT_LF 0x1
19 #define CONVERT_STAT_BITS_TXT_CRLF 0x2
20 #define CONVERT_STAT_BITS_BIN 0x4
21
22 enum crlf_action {
23 CRLF_UNDEFINED,
24 CRLF_BINARY,
25 CRLF_TEXT,
26 CRLF_TEXT_INPUT,
27 CRLF_TEXT_CRLF,
28 CRLF_AUTO,
29 CRLF_AUTO_INPUT,
30 CRLF_AUTO_CRLF
31 };
32
33 struct text_stat {
34 /* NUL, CR, LF and CRLF counts */
35 unsigned nul, lonecr, lonelf, crlf;
36
37 /* These are just approximations! */
38 unsigned printable, nonprintable;
39 };
40
41 static void gather_stats(const char *buf, unsigned long size, struct text_stat *stats)
42 {
43 unsigned long i;
44
45 memset(stats, 0, sizeof(*stats));
46
47 for (i = 0; i < size; i++) {
48 unsigned char c = buf[i];
49 if (c == '\r') {
50 if (i+1 < size && buf[i+1] == '\n') {
51 stats->crlf++;
52 i++;
53 } else
54 stats->lonecr++;
55 continue;
56 }
57 if (c == '\n') {
58 stats->lonelf++;
59 continue;
60 }
61 if (c == 127)
62 /* DEL */
63 stats->nonprintable++;
64 else if (c < 32) {
65 switch (c) {
66 /* BS, HT, ESC and FF */
67 case '\b': case '\t': case '\033': case '\014':
68 stats->printable++;
69 break;
70 case 0:
71 stats->nul++;
72 /* fall through */
73 default:
74 stats->nonprintable++;
75 }
76 }
77 else
78 stats->printable++;
79 }
80
81 /* If file ends with EOF then don't count this EOF as non-printable. */
82 if (size >= 1 && buf[size-1] == '\032')
83 stats->nonprintable--;
84 }
85
86 /*
87 * The same heuristics as diff.c::mmfile_is_binary()
88 * We treat files with bare CR as binary
89 */
90 static int convert_is_binary(unsigned long size, const struct text_stat *stats)
91 {
92 if (stats->lonecr)
93 return 1;
94 if (stats->nul)
95 return 1;
96 if ((stats->printable >> 7) < stats->nonprintable)
97 return 1;
98 return 0;
99 }
100
101 static unsigned int gather_convert_stats(const char *data, unsigned long size)
102 {
103 struct text_stat stats;
104 int ret = 0;
105 if (!data || !size)
106 return 0;
107 gather_stats(data, size, &stats);
108 if (convert_is_binary(size, &stats))
109 ret |= CONVERT_STAT_BITS_BIN;
110 if (stats.crlf)
111 ret |= CONVERT_STAT_BITS_TXT_CRLF;
112 if (stats.lonelf)
113 ret |= CONVERT_STAT_BITS_TXT_LF;
114
115 return ret;
116 }
117
118 static const char *gather_convert_stats_ascii(const char *data, unsigned long size)
119 {
120 unsigned int convert_stats = gather_convert_stats(data, size);
121
122 if (convert_stats & CONVERT_STAT_BITS_BIN)
123 return "-text";
124 switch (convert_stats) {
125 case CONVERT_STAT_BITS_TXT_LF:
126 return "lf";
127 case CONVERT_STAT_BITS_TXT_CRLF:
128 return "crlf";
129 case CONVERT_STAT_BITS_TXT_LF | CONVERT_STAT_BITS_TXT_CRLF:
130 return "mixed";
131 default:
132 return "none";
133 }
134 }
135
136 const char *get_cached_convert_stats_ascii(const char *path)
137 {
138 const char *ret;
139 unsigned long sz;
140 void *data = read_blob_data_from_cache(path, &sz);
141 ret = gather_convert_stats_ascii(data, sz);
142 free(data);
143 return ret;
144 }
145
146 const char *get_wt_convert_stats_ascii(const char *path)
147 {
148 const char *ret = "";
149 struct strbuf sb = STRBUF_INIT;
150 if (strbuf_read_file(&sb, path, 0) >= 0)
151 ret = gather_convert_stats_ascii(sb.buf, sb.len);
152 strbuf_release(&sb);
153 return ret;
154 }
155
156 static int text_eol_is_crlf(void)
157 {
158 if (auto_crlf == AUTO_CRLF_TRUE)
159 return 1;
160 else if (auto_crlf == AUTO_CRLF_INPUT)
161 return 0;
162 if (core_eol == EOL_CRLF)
163 return 1;
164 if (core_eol == EOL_UNSET && EOL_NATIVE == EOL_CRLF)
165 return 1;
166 return 0;
167 }
168
169 static enum eol output_eol(enum crlf_action crlf_action)
170 {
171 switch (crlf_action) {
172 case CRLF_BINARY:
173 return EOL_UNSET;
174 case CRLF_TEXT_CRLF:
175 return EOL_CRLF;
176 case CRLF_TEXT_INPUT:
177 return EOL_LF;
178 case CRLF_UNDEFINED:
179 case CRLF_AUTO_CRLF:
180 return EOL_CRLF;
181 case CRLF_AUTO_INPUT:
182 return EOL_LF;
183 case CRLF_TEXT:
184 case CRLF_AUTO:
185 /* fall through */
186 return text_eol_is_crlf() ? EOL_CRLF : EOL_LF;
187 }
188 warning("Illegal crlf_action %d\n", (int)crlf_action);
189 return core_eol;
190 }
191
192 static void check_safe_crlf(const char *path, enum crlf_action crlf_action,
193 struct text_stat *old_stats, struct text_stat *new_stats,
194 enum safe_crlf checksafe)
195 {
196 if (old_stats->crlf && !new_stats->crlf ) {
197 /*
198 * CRLFs would not be restored by checkout
199 */
200 if (checksafe == SAFE_CRLF_WARN)
201 warning("CRLF will be replaced by LF in %s.\nThe file will have its original line endings in your working directory.", path);
202 else /* i.e. SAFE_CRLF_FAIL */
203 die("CRLF would be replaced by LF in %s.", path);
204 } else if (old_stats->lonelf && !new_stats->lonelf ) {
205 /*
206 * CRLFs would be added by checkout
207 */
208 if (checksafe == SAFE_CRLF_WARN)
209 warning("LF will be replaced by CRLF in %s.\nThe file will have its original line endings in your working directory.", path);
210 else /* i.e. SAFE_CRLF_FAIL */
211 die("LF would be replaced by CRLF in %s", path);
212 }
213 }
214
215 static int has_cr_in_index(const char *path)
216 {
217 unsigned long sz;
218 void *data;
219 int has_cr;
220
221 data = read_blob_data_from_cache(path, &sz);
222 if (!data)
223 return 0;
224 has_cr = memchr(data, '\r', sz) != NULL;
225 free(data);
226 return has_cr;
227 }
228
229 static int will_convert_lf_to_crlf(size_t len, struct text_stat *stats,
230 enum crlf_action crlf_action)
231 {
232 if (output_eol(crlf_action) != EOL_CRLF)
233 return 0;
234 /* No "naked" LF? Nothing to convert, regardless. */
235 if (!stats->lonelf)
236 return 0;
237
238 if (crlf_action == CRLF_AUTO || crlf_action == CRLF_AUTO_INPUT || crlf_action == CRLF_AUTO_CRLF) {
239 /* If we have any CR or CRLF line endings, we do not touch it */
240 /* This is the new safer autocrlf-handling */
241 if (stats->lonecr || stats->crlf)
242 return 0;
243
244 if (convert_is_binary(len, stats))
245 return 0;
246 }
247 return 1;
248
249 }
250
251 static int crlf_to_git(const char *path, const char *src, size_t len,
252 struct strbuf *buf,
253 enum crlf_action crlf_action, enum safe_crlf checksafe)
254 {
255 struct text_stat stats;
256 char *dst;
257 int convert_crlf_into_lf;
258
259 if (crlf_action == CRLF_BINARY ||
260 (src && !len))
261 return 0;
262
263 /*
264 * If we are doing a dry-run and have no source buffer, there is
265 * nothing to analyze; we must assume we would convert.
266 */
267 if (!buf && !src)
268 return 1;
269
270 gather_stats(src, len, &stats);
271 /* Optimization: No CRLF? Nothing to convert, regardless. */
272 convert_crlf_into_lf = !!stats.crlf;
273
274 if (crlf_action == CRLF_AUTO || crlf_action == CRLF_AUTO_INPUT || crlf_action == CRLF_AUTO_CRLF) {
275 if (convert_is_binary(len, &stats))
276 return 0;
277 /*
278 * If the file in the index has any CR in it, do not convert.
279 * This is the new safer autocrlf handling.
280 */
281 if (checksafe == SAFE_CRLF_RENORMALIZE)
282 checksafe = SAFE_CRLF_FALSE;
283 else if (has_cr_in_index(path))
284 convert_crlf_into_lf = 0;
285 }
286 if (checksafe && len) {
287 struct text_stat new_stats;
288 memcpy(&new_stats, &stats, sizeof(new_stats));
289 /* simulate "git add" */
290 if (convert_crlf_into_lf) {
291 new_stats.lonelf += new_stats.crlf;
292 new_stats.crlf = 0;
293 }
294 /* simulate "git checkout" */
295 if (will_convert_lf_to_crlf(len, &new_stats, crlf_action)) {
296 new_stats.crlf += new_stats.lonelf;
297 new_stats.lonelf = 0;
298 }
299 check_safe_crlf(path, crlf_action, &stats, &new_stats, checksafe);
300 }
301 if (!convert_crlf_into_lf)
302 return 0;
303
304 /*
305 * At this point all of our source analysis is done, and we are sure we
306 * would convert. If we are in dry-run mode, we can give an answer.
307 */
308 if (!buf)
309 return 1;
310
311 /* only grow if not in place */
312 if (strbuf_avail(buf) + buf->len < len)
313 strbuf_grow(buf, len - buf->len);
314 dst = buf->buf;
315 if (crlf_action == CRLF_AUTO || crlf_action == CRLF_AUTO_INPUT || crlf_action == CRLF_AUTO_CRLF) {
316 /*
317 * If we guessed, we already know we rejected a file with
318 * lone CR, and we can strip a CR without looking at what
319 * follow it.
320 */
321 do {
322 unsigned char c = *src++;
323 if (c != '\r')
324 *dst++ = c;
325 } while (--len);
326 } else {
327 do {
328 unsigned char c = *src++;
329 if (! (c == '\r' && (1 < len && *src == '\n')))
330 *dst++ = c;
331 } while (--len);
332 }
333 strbuf_setlen(buf, dst - buf->buf);
334 return 1;
335 }
336
337 static int crlf_to_worktree(const char *path, const char *src, size_t len,
338 struct strbuf *buf, enum crlf_action crlf_action)
339 {
340 char *to_free = NULL;
341 struct text_stat stats;
342
343 if (!len || output_eol(crlf_action) != EOL_CRLF)
344 return 0;
345
346 gather_stats(src, len, &stats);
347 if (!will_convert_lf_to_crlf(len, &stats, crlf_action))
348 return 0;
349
350 /* are we "faking" in place editing ? */
351 if (src == buf->buf)
352 to_free = strbuf_detach(buf, NULL);
353
354 strbuf_grow(buf, len + stats.lonelf);
355 for (;;) {
356 const char *nl = memchr(src, '\n', len);
357 if (!nl)
358 break;
359 if (nl > src && nl[-1] == '\r') {
360 strbuf_add(buf, src, nl + 1 - src);
361 } else {
362 strbuf_add(buf, src, nl - src);
363 strbuf_addstr(buf, "\r\n");
364 }
365 len -= nl + 1 - src;
366 src = nl + 1;
367 }
368 strbuf_add(buf, src, len);
369
370 free(to_free);
371 return 1;
372 }
373
374 struct filter_params {
375 const char *src;
376 unsigned long size;
377 int fd;
378 const char *cmd;
379 const char *path;
380 };
381
382 static int filter_buffer_or_fd(int in, int out, void *data)
383 {
384 /*
385 * Spawn cmd and feed the buffer contents through its stdin.
386 */
387 struct child_process child_process = CHILD_PROCESS_INIT;
388 struct filter_params *params = (struct filter_params *)data;
389 int write_err, status;
390 const char *argv[] = { NULL, NULL };
391
392 /* apply % substitution to cmd */
393 struct strbuf cmd = STRBUF_INIT;
394 struct strbuf path = STRBUF_INIT;
395 struct strbuf_expand_dict_entry dict[] = {
396 { "f", NULL, },
397 { NULL, NULL, },
398 };
399
400 /* quote the path to preserve spaces, etc. */
401 sq_quote_buf(&path, params->path);
402 dict[0].value = path.buf;
403
404 /* expand all %f with the quoted path */
405 strbuf_expand(&cmd, params->cmd, strbuf_expand_dict_cb, &dict);
406 strbuf_release(&path);
407
408 argv[0] = cmd.buf;
409
410 child_process.argv = argv;
411 child_process.use_shell = 1;
412 child_process.in = -1;
413 child_process.out = out;
414
415 if (start_command(&child_process))
416 return error("cannot fork to run external filter '%s'", params->cmd);
417
418 sigchain_push(SIGPIPE, SIG_IGN);
419
420 if (params->src) {
421 write_err = (write_in_full(child_process.in,
422 params->src, params->size) < 0);
423 if (errno == EPIPE)
424 write_err = 0;
425 } else {
426 write_err = copy_fd(params->fd, child_process.in);
427 if (write_err == COPY_WRITE_ERROR && errno == EPIPE)
428 write_err = 0;
429 }
430
431 if (close(child_process.in))
432 write_err = 1;
433 if (write_err)
434 error("cannot feed the input to external filter '%s'", params->cmd);
435
436 sigchain_pop(SIGPIPE);
437
438 status = finish_command(&child_process);
439 if (status)
440 error("external filter '%s' failed %d", params->cmd, status);
441
442 strbuf_release(&cmd);
443 return (write_err || status);
444 }
445
446 static int apply_single_file_filter(const char *path, const char *src, size_t len, int fd,
447 struct strbuf *dst, const char *cmd)
448 {
449 /*
450 * Create a pipeline to have the command filter the buffer's
451 * contents.
452 *
453 * (child --> cmd) --> us
454 */
455 int err = 0;
456 struct strbuf nbuf = STRBUF_INIT;
457 struct async async;
458 struct filter_params params;
459
460 memset(&async, 0, sizeof(async));
461 async.proc = filter_buffer_or_fd;
462 async.data = &params;
463 async.out = -1;
464 params.src = src;
465 params.size = len;
466 params.fd = fd;
467 params.cmd = cmd;
468 params.path = path;
469
470 fflush(NULL);
471 if (start_async(&async))
472 return 0; /* error was already reported */
473
474 if (strbuf_read(&nbuf, async.out, len) < 0) {
475 err = error("read from external filter '%s' failed", cmd);
476 }
477 if (close(async.out)) {
478 err = error("read from external filter '%s' failed", cmd);
479 }
480 if (finish_async(&async)) {
481 err = error("external filter '%s' failed", cmd);
482 }
483
484 if (!err) {
485 strbuf_swap(dst, &nbuf);
486 }
487 strbuf_release(&nbuf);
488 return !err;
489 }
490
491 #define CAP_CLEAN (1u<<0)
492 #define CAP_SMUDGE (1u<<1)
493
494 struct cmd2process {
495 struct hashmap_entry ent; /* must be the first member! */
496 unsigned int supported_capabilities;
497 const char *cmd;
498 struct child_process process;
499 };
500
501 static int cmd_process_map_initialized;
502 static struct hashmap cmd_process_map;
503
504 static int cmd2process_cmp(const struct cmd2process *e1,
505 const struct cmd2process *e2,
506 const void *unused)
507 {
508 return strcmp(e1->cmd, e2->cmd);
509 }
510
511 static struct cmd2process *find_multi_file_filter_entry(struct hashmap *hashmap, const char *cmd)
512 {
513 struct cmd2process key;
514 hashmap_entry_init(&key, strhash(cmd));
515 key.cmd = cmd;
516 return hashmap_get(hashmap, &key, NULL);
517 }
518
519 static int packet_write_list(int fd, const char *line, ...)
520 {
521 va_list args;
522 int err;
523 va_start(args, line);
524 for (;;) {
525 if (!line)
526 break;
527 if (strlen(line) > LARGE_PACKET_DATA_MAX)
528 return -1;
529 err = packet_write_fmt_gently(fd, "%s\n", line);
530 if (err)
531 return err;
532 line = va_arg(args, const char*);
533 }
534 va_end(args);
535 return packet_flush_gently(fd);
536 }
537
538 static void read_multi_file_filter_status(int fd, struct strbuf *status)
539 {
540 struct strbuf **pair;
541 char *line;
542 for (;;) {
543 line = packet_read_line(fd, NULL);
544 if (!line)
545 break;
546 pair = strbuf_split_str(line, '=', 2);
547 if (pair[0] && pair[0]->len && pair[1]) {
548 /* the last "status=<foo>" line wins */
549 if (!strcmp(pair[0]->buf, "status=")) {
550 strbuf_reset(status);
551 strbuf_addbuf(status, pair[1]);
552 }
553 }
554 strbuf_list_free(pair);
555 }
556 }
557
558 static void kill_multi_file_filter(struct hashmap *hashmap, struct cmd2process *entry)
559 {
560 if (!entry)
561 return;
562
563 entry->process.clean_on_exit = 0;
564 kill(entry->process.pid, SIGTERM);
565 finish_command(&entry->process);
566
567 hashmap_remove(hashmap, entry, NULL);
568 free(entry);
569 }
570
571 static void stop_multi_file_filter(struct child_process *process)
572 {
573 sigchain_push(SIGPIPE, SIG_IGN);
574 /* Closing the pipe signals the filter to initiate a shutdown. */
575 close(process->in);
576 close(process->out);
577 sigchain_pop(SIGPIPE);
578 /* Finish command will wait until the shutdown is complete. */
579 finish_command(process);
580 }
581
582 static struct cmd2process *start_multi_file_filter(struct hashmap *hashmap, const char *cmd)
583 {
584 int err;
585 struct cmd2process *entry;
586 struct child_process *process;
587 const char *argv[] = { cmd, NULL };
588 struct string_list cap_list = STRING_LIST_INIT_NODUP;
589 char *cap_buf;
590 const char *cap_name;
591
592 entry = xmalloc(sizeof(*entry));
593 entry->cmd = cmd;
594 entry->supported_capabilities = 0;
595 process = &entry->process;
596
597 child_process_init(process);
598 process->argv = argv;
599 process->use_shell = 1;
600 process->in = -1;
601 process->out = -1;
602 process->clean_on_exit = 1;
603 process->clean_on_exit_handler = stop_multi_file_filter;
604
605 if (start_command(process)) {
606 error("cannot fork to run external filter '%s'", cmd);
607 return NULL;
608 }
609
610 hashmap_entry_init(entry, strhash(cmd));
611
612 sigchain_push(SIGPIPE, SIG_IGN);
613
614 err = packet_write_list(process->in, "git-filter-client", "version=2", NULL);
615 if (err)
616 goto done;
617
618 err = strcmp(packet_read_line(process->out, NULL), "git-filter-server");
619 if (err) {
620 error("external filter '%s' does not support filter protocol version 2", cmd);
621 goto done;
622 }
623 err = strcmp(packet_read_line(process->out, NULL), "version=2");
624 if (err)
625 goto done;
626 err = packet_read_line(process->out, NULL) != NULL;
627 if (err)
628 goto done;
629
630 err = packet_write_list(process->in, "capability=clean", "capability=smudge", NULL);
631
632 for (;;) {
633 cap_buf = packet_read_line(process->out, NULL);
634 if (!cap_buf)
635 break;
636 string_list_split_in_place(&cap_list, cap_buf, '=', 1);
637
638 if (cap_list.nr != 2 || strcmp(cap_list.items[0].string, "capability"))
639 continue;
640
641 cap_name = cap_list.items[1].string;
642 if (!strcmp(cap_name, "clean")) {
643 entry->supported_capabilities |= CAP_CLEAN;
644 } else if (!strcmp(cap_name, "smudge")) {
645 entry->supported_capabilities |= CAP_SMUDGE;
646 } else {
647 warning(
648 "external filter '%s' requested unsupported filter capability '%s'",
649 cmd, cap_name
650 );
651 }
652
653 string_list_clear(&cap_list, 0);
654 }
655
656 done:
657 sigchain_pop(SIGPIPE);
658
659 if (err || errno == EPIPE) {
660 error("initialization for external filter '%s' failed", cmd);
661 kill_multi_file_filter(hashmap, entry);
662 return NULL;
663 }
664
665 hashmap_add(hashmap, entry);
666 return entry;
667 }
668
669 static int apply_multi_file_filter(const char *path, const char *src, size_t len,
670 int fd, struct strbuf *dst, const char *cmd,
671 const unsigned int wanted_capability)
672 {
673 int err;
674 struct cmd2process *entry;
675 struct child_process *process;
676 struct strbuf nbuf = STRBUF_INIT;
677 struct strbuf filter_status = STRBUF_INIT;
678 const char *filter_type;
679
680 if (!cmd_process_map_initialized) {
681 cmd_process_map_initialized = 1;
682 hashmap_init(&cmd_process_map, (hashmap_cmp_fn) cmd2process_cmp, 0);
683 entry = NULL;
684 } else {
685 entry = find_multi_file_filter_entry(&cmd_process_map, cmd);
686 }
687
688 fflush(NULL);
689
690 if (!entry) {
691 entry = start_multi_file_filter(&cmd_process_map, cmd);
692 if (!entry)
693 return 0;
694 }
695 process = &entry->process;
696
697 if (!(wanted_capability & entry->supported_capabilities))
698 return 0;
699
700 if (CAP_CLEAN & wanted_capability)
701 filter_type = "clean";
702 else if (CAP_SMUDGE & wanted_capability)
703 filter_type = "smudge";
704 else
705 die("unexpected filter type");
706
707 sigchain_push(SIGPIPE, SIG_IGN);
708
709 assert(strlen(filter_type) < LARGE_PACKET_DATA_MAX - strlen("command=\n"));
710 err = packet_write_fmt_gently(process->in, "command=%s\n", filter_type);
711 if (err)
712 goto done;
713
714 err = strlen(path) > LARGE_PACKET_DATA_MAX - strlen("pathname=\n");
715 if (err) {
716 error("path name too long for external filter");
717 goto done;
718 }
719
720 err = packet_write_fmt_gently(process->in, "pathname=%s\n", path);
721 if (err)
722 goto done;
723
724 err = packet_flush_gently(process->in);
725 if (err)
726 goto done;
727
728 if (fd >= 0)
729 err = write_packetized_from_fd(fd, process->in);
730 else
731 err = write_packetized_from_buf(src, len, process->in);
732 if (err)
733 goto done;
734
735 read_multi_file_filter_status(process->out, &filter_status);
736 err = strcmp(filter_status.buf, "success");
737 if (err)
738 goto done;
739
740 err = read_packetized_to_strbuf(process->out, &nbuf) < 0;
741 if (err)
742 goto done;
743
744 read_multi_file_filter_status(process->out, &filter_status);
745 err = strcmp(filter_status.buf, "success");
746
747 done:
748 sigchain_pop(SIGPIPE);
749
750 if (err || errno == EPIPE) {
751 if (!strcmp(filter_status.buf, "error")) {
752 /* The filter signaled a problem with the file. */
753 } else if (!strcmp(filter_status.buf, "abort")) {
754 /*
755 * The filter signaled a permanent problem. Don't try to filter
756 * files with the same command for the lifetime of the current
757 * Git process.
758 */
759 entry->supported_capabilities &= ~wanted_capability;
760 } else {
761 /*
762 * Something went wrong with the protocol filter.
763 * Force shutdown and restart if another blob requires filtering.
764 */
765 error("external filter '%s' failed", cmd);
766 kill_multi_file_filter(&cmd_process_map, entry);
767 }
768 } else {
769 strbuf_swap(dst, &nbuf);
770 }
771 strbuf_release(&nbuf);
772 return !err;
773 }
774
775 static struct convert_driver {
776 const char *name;
777 struct convert_driver *next;
778 const char *smudge;
779 const char *clean;
780 const char *process;
781 int required;
782 } *user_convert, **user_convert_tail;
783
784 static int apply_filter(const char *path, const char *src, size_t len,
785 int fd, struct strbuf *dst, struct convert_driver *drv,
786 const unsigned int wanted_capability)
787 {
788 const char *cmd = NULL;
789
790 if (!drv)
791 return 0;
792
793 if (!dst)
794 return 1;
795
796 if ((CAP_CLEAN & wanted_capability) && !drv->process && drv->clean)
797 cmd = drv->clean;
798 else if ((CAP_SMUDGE & wanted_capability) && !drv->process && drv->smudge)
799 cmd = drv->smudge;
800
801 if (cmd && *cmd)
802 return apply_single_file_filter(path, src, len, fd, dst, cmd);
803 else if (drv->process && *drv->process)
804 return apply_multi_file_filter(path, src, len, fd, dst, drv->process, wanted_capability);
805
806 return 0;
807 }
808
809 static int read_convert_config(const char *var, const char *value, void *cb)
810 {
811 const char *key, *name;
812 int namelen;
813 struct convert_driver *drv;
814
815 /*
816 * External conversion drivers are configured using
817 * "filter.<name>.variable".
818 */
819 if (parse_config_key(var, "filter", &name, &namelen, &key) < 0 || !name)
820 return 0;
821 for (drv = user_convert; drv; drv = drv->next)
822 if (!strncmp(drv->name, name, namelen) && !drv->name[namelen])
823 break;
824 if (!drv) {
825 drv = xcalloc(1, sizeof(struct convert_driver));
826 drv->name = xmemdupz(name, namelen);
827 *user_convert_tail = drv;
828 user_convert_tail = &(drv->next);
829 }
830
831 /*
832 * filter.<name>.smudge and filter.<name>.clean specifies
833 * the command line:
834 *
835 * command-line
836 *
837 * The command-line will not be interpolated in any way.
838 */
839
840 if (!strcmp("smudge", key))
841 return git_config_string(&drv->smudge, var, value);
842
843 if (!strcmp("clean", key))
844 return git_config_string(&drv->clean, var, value);
845
846 if (!strcmp("process", key))
847 return git_config_string(&drv->process, var, value);
848
849 if (!strcmp("required", key)) {
850 drv->required = git_config_bool(var, value);
851 return 0;
852 }
853
854 return 0;
855 }
856
857 static int count_ident(const char *cp, unsigned long size)
858 {
859 /*
860 * "$Id: 0000000000000000000000000000000000000000 $" <=> "$Id$"
861 */
862 int cnt = 0;
863 char ch;
864
865 while (size) {
866 ch = *cp++;
867 size--;
868 if (ch != '$')
869 continue;
870 if (size < 3)
871 break;
872 if (memcmp("Id", cp, 2))
873 continue;
874 ch = cp[2];
875 cp += 3;
876 size -= 3;
877 if (ch == '$')
878 cnt++; /* $Id$ */
879 if (ch != ':')
880 continue;
881
882 /*
883 * "$Id: ... "; scan up to the closing dollar sign and discard.
884 */
885 while (size) {
886 ch = *cp++;
887 size--;
888 if (ch == '$') {
889 cnt++;
890 break;
891 }
892 if (ch == '\n')
893 break;
894 }
895 }
896 return cnt;
897 }
898
899 static int ident_to_git(const char *path, const char *src, size_t len,
900 struct strbuf *buf, int ident)
901 {
902 char *dst, *dollar;
903
904 if (!ident || (src && !count_ident(src, len)))
905 return 0;
906
907 if (!buf)
908 return 1;
909
910 /* only grow if not in place */
911 if (strbuf_avail(buf) + buf->len < len)
912 strbuf_grow(buf, len - buf->len);
913 dst = buf->buf;
914 for (;;) {
915 dollar = memchr(src, '$', len);
916 if (!dollar)
917 break;
918 memmove(dst, src, dollar + 1 - src);
919 dst += dollar + 1 - src;
920 len -= dollar + 1 - src;
921 src = dollar + 1;
922
923 if (len > 3 && !memcmp(src, "Id:", 3)) {
924 dollar = memchr(src + 3, '$', len - 3);
925 if (!dollar)
926 break;
927 if (memchr(src + 3, '\n', dollar - src - 3)) {
928 /* Line break before the next dollar. */
929 continue;
930 }
931
932 memcpy(dst, "Id$", 3);
933 dst += 3;
934 len -= dollar + 1 - src;
935 src = dollar + 1;
936 }
937 }
938 memmove(dst, src, len);
939 strbuf_setlen(buf, dst + len - buf->buf);
940 return 1;
941 }
942
943 static int ident_to_worktree(const char *path, const char *src, size_t len,
944 struct strbuf *buf, int ident)
945 {
946 unsigned char sha1[20];
947 char *to_free = NULL, *dollar, *spc;
948 int cnt;
949
950 if (!ident)
951 return 0;
952
953 cnt = count_ident(src, len);
954 if (!cnt)
955 return 0;
956
957 /* are we "faking" in place editing ? */
958 if (src == buf->buf)
959 to_free = strbuf_detach(buf, NULL);
960 hash_sha1_file(src, len, "blob", sha1);
961
962 strbuf_grow(buf, len + cnt * 43);
963 for (;;) {
964 /* step 1: run to the next '$' */
965 dollar = memchr(src, '$', len);
966 if (!dollar)
967 break;
968 strbuf_add(buf, src, dollar + 1 - src);
969 len -= dollar + 1 - src;
970 src = dollar + 1;
971
972 /* step 2: does it looks like a bit like Id:xxx$ or Id$ ? */
973 if (len < 3 || memcmp("Id", src, 2))
974 continue;
975
976 /* step 3: skip over Id$ or Id:xxxxx$ */
977 if (src[2] == '$') {
978 src += 3;
979 len -= 3;
980 } else if (src[2] == ':') {
981 /*
982 * It's possible that an expanded Id has crept its way into the
983 * repository, we cope with that by stripping the expansion out.
984 * This is probably not a good idea, since it will cause changes
985 * on checkout, which won't go away by stash, but let's keep it
986 * for git-style ids.
987 */
988 dollar = memchr(src + 3, '$', len - 3);
989 if (!dollar) {
990 /* incomplete keyword, no more '$', so just quit the loop */
991 break;
992 }
993
994 if (memchr(src + 3, '\n', dollar - src - 3)) {
995 /* Line break before the next dollar. */
996 continue;
997 }
998
999 spc = memchr(src + 4, ' ', dollar - src - 4);
1000 if (spc && spc < dollar-1) {
1001 /* There are spaces in unexpected places.
1002 * This is probably an id from some other
1003 * versioning system. Keep it for now.
1004 */
1005 continue;
1006 }
1007
1008 len -= dollar + 1 - src;
1009 src = dollar + 1;
1010 } else {
1011 /* it wasn't a "Id$" or "Id:xxxx$" */
1012 continue;
1013 }
1014
1015 /* step 4: substitute */
1016 strbuf_addstr(buf, "Id: ");
1017 strbuf_add(buf, sha1_to_hex(sha1), 40);
1018 strbuf_addstr(buf, " $");
1019 }
1020 strbuf_add(buf, src, len);
1021
1022 free(to_free);
1023 return 1;
1024 }
1025
1026 static enum crlf_action git_path_check_crlf(struct git_attr_check *check)
1027 {
1028 const char *value = check->value;
1029
1030 if (ATTR_TRUE(value))
1031 return CRLF_TEXT;
1032 else if (ATTR_FALSE(value))
1033 return CRLF_BINARY;
1034 else if (ATTR_UNSET(value))
1035 ;
1036 else if (!strcmp(value, "input"))
1037 return CRLF_TEXT_INPUT;
1038 else if (!strcmp(value, "auto"))
1039 return CRLF_AUTO;
1040 return CRLF_UNDEFINED;
1041 }
1042
1043 static enum eol git_path_check_eol(struct git_attr_check *check)
1044 {
1045 const char *value = check->value;
1046
1047 if (ATTR_UNSET(value))
1048 ;
1049 else if (!strcmp(value, "lf"))
1050 return EOL_LF;
1051 else if (!strcmp(value, "crlf"))
1052 return EOL_CRLF;
1053 return EOL_UNSET;
1054 }
1055
1056 static struct convert_driver *git_path_check_convert(struct git_attr_check *check)
1057 {
1058 const char *value = check->value;
1059 struct convert_driver *drv;
1060
1061 if (ATTR_TRUE(value) || ATTR_FALSE(value) || ATTR_UNSET(value))
1062 return NULL;
1063 for (drv = user_convert; drv; drv = drv->next)
1064 if (!strcmp(value, drv->name))
1065 return drv;
1066 return NULL;
1067 }
1068
1069 static int git_path_check_ident(struct git_attr_check *check)
1070 {
1071 const char *value = check->value;
1072
1073 return !!ATTR_TRUE(value);
1074 }
1075
1076 struct conv_attrs {
1077 struct convert_driver *drv;
1078 enum crlf_action attr_action; /* What attr says */
1079 enum crlf_action crlf_action; /* When no attr is set, use core.autocrlf */
1080 int ident;
1081 };
1082
1083 static const char *conv_attr_name[] = {
1084 "crlf", "ident", "filter", "eol", "text",
1085 };
1086 #define NUM_CONV_ATTRS ARRAY_SIZE(conv_attr_name)
1087
1088 static void convert_attrs(struct conv_attrs *ca, const char *path)
1089 {
1090 int i;
1091 static struct git_attr_check ccheck[NUM_CONV_ATTRS];
1092
1093 if (!ccheck[0].attr) {
1094 for (i = 0; i < NUM_CONV_ATTRS; i++)
1095 ccheck[i].attr = git_attr(conv_attr_name[i]);
1096 user_convert_tail = &user_convert;
1097 git_config(read_convert_config, NULL);
1098 }
1099
1100 if (!git_check_attr(path, NUM_CONV_ATTRS, ccheck)) {
1101 ca->crlf_action = git_path_check_crlf(ccheck + 4);
1102 if (ca->crlf_action == CRLF_UNDEFINED)
1103 ca->crlf_action = git_path_check_crlf(ccheck + 0);
1104 ca->attr_action = ca->crlf_action;
1105 ca->ident = git_path_check_ident(ccheck + 1);
1106 ca->drv = git_path_check_convert(ccheck + 2);
1107 if (ca->crlf_action != CRLF_BINARY) {
1108 enum eol eol_attr = git_path_check_eol(ccheck + 3);
1109 if (ca->crlf_action == CRLF_AUTO && eol_attr == EOL_LF)
1110 ca->crlf_action = CRLF_AUTO_INPUT;
1111 else if (ca->crlf_action == CRLF_AUTO && eol_attr == EOL_CRLF)
1112 ca->crlf_action = CRLF_AUTO_CRLF;
1113 else if (eol_attr == EOL_LF)
1114 ca->crlf_action = CRLF_TEXT_INPUT;
1115 else if (eol_attr == EOL_CRLF)
1116 ca->crlf_action = CRLF_TEXT_CRLF;
1117 }
1118 ca->attr_action = ca->crlf_action;
1119 } else {
1120 ca->drv = NULL;
1121 ca->crlf_action = CRLF_UNDEFINED;
1122 ca->ident = 0;
1123 }
1124 if (ca->crlf_action == CRLF_TEXT)
1125 ca->crlf_action = text_eol_is_crlf() ? CRLF_TEXT_CRLF : CRLF_TEXT_INPUT;
1126 if (ca->crlf_action == CRLF_UNDEFINED && auto_crlf == AUTO_CRLF_FALSE)
1127 ca->crlf_action = CRLF_BINARY;
1128 if (ca->crlf_action == CRLF_UNDEFINED && auto_crlf == AUTO_CRLF_TRUE)
1129 ca->crlf_action = CRLF_AUTO_CRLF;
1130 if (ca->crlf_action == CRLF_UNDEFINED && auto_crlf == AUTO_CRLF_INPUT)
1131 ca->crlf_action = CRLF_AUTO_INPUT;
1132 }
1133
1134 int would_convert_to_git_filter_fd(const char *path)
1135 {
1136 struct conv_attrs ca;
1137
1138 convert_attrs(&ca, path);
1139 if (!ca.drv)
1140 return 0;
1141
1142 /*
1143 * Apply a filter to an fd only if the filter is required to succeed.
1144 * We must die if the filter fails, because the original data before
1145 * filtering is not available.
1146 */
1147 if (!ca.drv->required)
1148 return 0;
1149
1150 return apply_filter(path, NULL, 0, -1, NULL, ca.drv, CAP_CLEAN);
1151 }
1152
1153 const char *get_convert_attr_ascii(const char *path)
1154 {
1155 struct conv_attrs ca;
1156
1157 convert_attrs(&ca, path);
1158 switch (ca.attr_action) {
1159 case CRLF_UNDEFINED:
1160 return "";
1161 case CRLF_BINARY:
1162 return "-text";
1163 case CRLF_TEXT:
1164 return "text";
1165 case CRLF_TEXT_INPUT:
1166 return "text eol=lf";
1167 case CRLF_TEXT_CRLF:
1168 return "text eol=crlf";
1169 case CRLF_AUTO:
1170 return "text=auto";
1171 case CRLF_AUTO_CRLF:
1172 return "text=auto eol=crlf";
1173 case CRLF_AUTO_INPUT:
1174 return "text=auto eol=lf";
1175 }
1176 return "";
1177 }
1178
1179 int convert_to_git(const char *path, const char *src, size_t len,
1180 struct strbuf *dst, enum safe_crlf checksafe)
1181 {
1182 int ret = 0;
1183 struct conv_attrs ca;
1184
1185 convert_attrs(&ca, path);
1186
1187 ret |= apply_filter(path, src, len, -1, dst, ca.drv, CAP_CLEAN);
1188 if (!ret && ca.drv && ca.drv->required)
1189 die("%s: clean filter '%s' failed", path, ca.drv->name);
1190
1191 if (ret && dst) {
1192 src = dst->buf;
1193 len = dst->len;
1194 }
1195 ret |= crlf_to_git(path, src, len, dst, ca.crlf_action, checksafe);
1196 if (ret && dst) {
1197 src = dst->buf;
1198 len = dst->len;
1199 }
1200 return ret | ident_to_git(path, src, len, dst, ca.ident);
1201 }
1202
1203 void convert_to_git_filter_fd(const char *path, int fd, struct strbuf *dst,
1204 enum safe_crlf checksafe)
1205 {
1206 struct conv_attrs ca;
1207 convert_attrs(&ca, path);
1208
1209 assert(ca.drv);
1210 assert(ca.drv->clean || ca.drv->process);
1211
1212 if (!apply_filter(path, NULL, 0, fd, dst, ca.drv, CAP_CLEAN))
1213 die("%s: clean filter '%s' failed", path, ca.drv->name);
1214
1215 crlf_to_git(path, dst->buf, dst->len, dst, ca.crlf_action, checksafe);
1216 ident_to_git(path, dst->buf, dst->len, dst, ca.ident);
1217 }
1218
1219 static int convert_to_working_tree_internal(const char *path, const char *src,
1220 size_t len, struct strbuf *dst,
1221 int normalizing)
1222 {
1223 int ret = 0, ret_filter = 0;
1224 struct conv_attrs ca;
1225
1226 convert_attrs(&ca, path);
1227
1228 ret |= ident_to_worktree(path, src, len, dst, ca.ident);
1229 if (ret) {
1230 src = dst->buf;
1231 len = dst->len;
1232 }
1233 /*
1234 * CRLF conversion can be skipped if normalizing, unless there
1235 * is a smudge or process filter (even if the process filter doesn't
1236 * support smudge). The filters might expect CRLFs.
1237 */
1238 if ((ca.drv && (ca.drv->smudge || ca.drv->process)) || !normalizing) {
1239 ret |= crlf_to_worktree(path, src, len, dst, ca.crlf_action);
1240 if (ret) {
1241 src = dst->buf;
1242 len = dst->len;
1243 }
1244 }
1245
1246 ret_filter = apply_filter(path, src, len, -1, dst, ca.drv, CAP_SMUDGE);
1247 if (!ret_filter && ca.drv && ca.drv->required)
1248 die("%s: smudge filter %s failed", path, ca.drv->name);
1249
1250 return ret | ret_filter;
1251 }
1252
1253 int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)
1254 {
1255 return convert_to_working_tree_internal(path, src, len, dst, 0);
1256 }
1257
1258 int renormalize_buffer(const char *path, const char *src, size_t len, struct strbuf *dst)
1259 {
1260 int ret = convert_to_working_tree_internal(path, src, len, dst, 1);
1261 if (ret) {
1262 src = dst->buf;
1263 len = dst->len;
1264 }
1265 return ret | convert_to_git(path, src, len, dst, SAFE_CRLF_RENORMALIZE);
1266 }
1267
1268 /*****************************************************************
1269 *
1270 * Streaming conversion support
1271 *
1272 *****************************************************************/
1273
1274 typedef int (*filter_fn)(struct stream_filter *,
1275 const char *input, size_t *isize_p,
1276 char *output, size_t *osize_p);
1277 typedef void (*free_fn)(struct stream_filter *);
1278
1279 struct stream_filter_vtbl {
1280 filter_fn filter;
1281 free_fn free;
1282 };
1283
1284 struct stream_filter {
1285 struct stream_filter_vtbl *vtbl;
1286 };
1287
1288 static int null_filter_fn(struct stream_filter *filter,
1289 const char *input, size_t *isize_p,
1290 char *output, size_t *osize_p)
1291 {
1292 size_t count;
1293
1294 if (!input)
1295 return 0; /* we do not keep any states */
1296 count = *isize_p;
1297 if (*osize_p < count)
1298 count = *osize_p;
1299 if (count) {
1300 memmove(output, input, count);
1301 *isize_p -= count;
1302 *osize_p -= count;
1303 }
1304 return 0;
1305 }
1306
1307 static void null_free_fn(struct stream_filter *filter)
1308 {
1309 ; /* nothing -- null instances are shared */
1310 }
1311
1312 static struct stream_filter_vtbl null_vtbl = {
1313 null_filter_fn,
1314 null_free_fn,
1315 };
1316
1317 static struct stream_filter null_filter_singleton = {
1318 &null_vtbl,
1319 };
1320
1321 int is_null_stream_filter(struct stream_filter *filter)
1322 {
1323 return filter == &null_filter_singleton;
1324 }
1325
1326
1327 /*
1328 * LF-to-CRLF filter
1329 */
1330
1331 struct lf_to_crlf_filter {
1332 struct stream_filter filter;
1333 unsigned has_held:1;
1334 char held;
1335 };
1336
1337 static int lf_to_crlf_filter_fn(struct stream_filter *filter,
1338 const char *input, size_t *isize_p,
1339 char *output, size_t *osize_p)
1340 {
1341 size_t count, o = 0;
1342 struct lf_to_crlf_filter *lf_to_crlf = (struct lf_to_crlf_filter *)filter;
1343
1344 /*
1345 * We may be holding onto the CR to see if it is followed by a
1346 * LF, in which case we would need to go to the main loop.
1347 * Otherwise, just emit it to the output stream.
1348 */
1349 if (lf_to_crlf->has_held && (lf_to_crlf->held != '\r' || !input)) {
1350 output[o++] = lf_to_crlf->held;
1351 lf_to_crlf->has_held = 0;
1352 }
1353
1354 /* We are told to drain */
1355 if (!input) {
1356 *osize_p -= o;
1357 return 0;
1358 }
1359
1360 count = *isize_p;
1361 if (count || lf_to_crlf->has_held) {
1362 size_t i;
1363 int was_cr = 0;
1364
1365 if (lf_to_crlf->has_held) {
1366 was_cr = 1;
1367 lf_to_crlf->has_held = 0;
1368 }
1369
1370 for (i = 0; o < *osize_p && i < count; i++) {
1371 char ch = input[i];
1372
1373 if (ch == '\n') {
1374 output[o++] = '\r';
1375 } else if (was_cr) {
1376 /*
1377 * Previous round saw CR and it is not followed
1378 * by a LF; emit the CR before processing the
1379 * current character.
1380 */
1381 output[o++] = '\r';
1382 }
1383
1384 /*
1385 * We may have consumed the last output slot,
1386 * in which case we need to break out of this
1387 * loop; hold the current character before
1388 * returning.
1389 */
1390 if (*osize_p <= o) {
1391 lf_to_crlf->has_held = 1;
1392 lf_to_crlf->held = ch;
1393 continue; /* break but increment i */
1394 }
1395
1396 if (ch == '\r') {
1397 was_cr = 1;
1398 continue;
1399 }
1400
1401 was_cr = 0;
1402 output[o++] = ch;
1403 }
1404
1405 *osize_p -= o;
1406 *isize_p -= i;
1407
1408 if (!lf_to_crlf->has_held && was_cr) {
1409 lf_to_crlf->has_held = 1;
1410 lf_to_crlf->held = '\r';
1411 }
1412 }
1413 return 0;
1414 }
1415
1416 static void lf_to_crlf_free_fn(struct stream_filter *filter)
1417 {
1418 free(filter);
1419 }
1420
1421 static struct stream_filter_vtbl lf_to_crlf_vtbl = {
1422 lf_to_crlf_filter_fn,
1423 lf_to_crlf_free_fn,
1424 };
1425
1426 static struct stream_filter *lf_to_crlf_filter(void)
1427 {
1428 struct lf_to_crlf_filter *lf_to_crlf = xcalloc(1, sizeof(*lf_to_crlf));
1429
1430 lf_to_crlf->filter.vtbl = &lf_to_crlf_vtbl;
1431 return (struct stream_filter *)lf_to_crlf;
1432 }
1433
1434 /*
1435 * Cascade filter
1436 */
1437 #define FILTER_BUFFER 1024
1438 struct cascade_filter {
1439 struct stream_filter filter;
1440 struct stream_filter *one;
1441 struct stream_filter *two;
1442 char buf[FILTER_BUFFER];
1443 int end, ptr;
1444 };
1445
1446 static int cascade_filter_fn(struct stream_filter *filter,
1447 const char *input, size_t *isize_p,
1448 char *output, size_t *osize_p)
1449 {
1450 struct cascade_filter *cas = (struct cascade_filter *) filter;
1451 size_t filled = 0;
1452 size_t sz = *osize_p;
1453 size_t to_feed, remaining;
1454
1455 /*
1456 * input -- (one) --> buf -- (two) --> output
1457 */
1458 while (filled < sz) {
1459 remaining = sz - filled;
1460
1461 /* do we already have something to feed two with? */
1462 if (cas->ptr < cas->end) {
1463 to_feed = cas->end - cas->ptr;
1464 if (stream_filter(cas->two,
1465 cas->buf + cas->ptr, &to_feed,
1466 output + filled, &remaining))
1467 return -1;
1468 cas->ptr += (cas->end - cas->ptr) - to_feed;
1469 filled = sz - remaining;
1470 continue;
1471 }
1472
1473 /* feed one from upstream and have it emit into our buffer */
1474 to_feed = input ? *isize_p : 0;
1475 if (input && !to_feed)
1476 break;
1477 remaining = sizeof(cas->buf);
1478 if (stream_filter(cas->one,
1479 input, &to_feed,
1480 cas->buf, &remaining))
1481 return -1;
1482 cas->end = sizeof(cas->buf) - remaining;
1483 cas->ptr = 0;
1484 if (input) {
1485 size_t fed = *isize_p - to_feed;
1486 *isize_p -= fed;
1487 input += fed;
1488 }
1489
1490 /* do we know that we drained one completely? */
1491 if (input || cas->end)
1492 continue;
1493
1494 /* tell two to drain; we have nothing more to give it */
1495 to_feed = 0;
1496 remaining = sz - filled;
1497 if (stream_filter(cas->two,
1498 NULL, &to_feed,
1499 output + filled, &remaining))
1500 return -1;
1501 if (remaining == (sz - filled))
1502 break; /* completely drained two */
1503 filled = sz - remaining;
1504 }
1505 *osize_p -= filled;
1506 return 0;
1507 }
1508
1509 static void cascade_free_fn(struct stream_filter *filter)
1510 {
1511 struct cascade_filter *cas = (struct cascade_filter *)filter;
1512 free_stream_filter(cas->one);
1513 free_stream_filter(cas->two);
1514 free(filter);
1515 }
1516
1517 static struct stream_filter_vtbl cascade_vtbl = {
1518 cascade_filter_fn,
1519 cascade_free_fn,
1520 };
1521
1522 static struct stream_filter *cascade_filter(struct stream_filter *one,
1523 struct stream_filter *two)
1524 {
1525 struct cascade_filter *cascade;
1526
1527 if (!one || is_null_stream_filter(one))
1528 return two;
1529 if (!two || is_null_stream_filter(two))
1530 return one;
1531
1532 cascade = xmalloc(sizeof(*cascade));
1533 cascade->one = one;
1534 cascade->two = two;
1535 cascade->end = cascade->ptr = 0;
1536 cascade->filter.vtbl = &cascade_vtbl;
1537 return (struct stream_filter *)cascade;
1538 }
1539
1540 /*
1541 * ident filter
1542 */
1543 #define IDENT_DRAINING (-1)
1544 #define IDENT_SKIPPING (-2)
1545 struct ident_filter {
1546 struct stream_filter filter;
1547 struct strbuf left;
1548 int state;
1549 char ident[45]; /* ": x40 $" */
1550 };
1551
1552 static int is_foreign_ident(const char *str)
1553 {
1554 int i;
1555
1556 if (!skip_prefix(str, "$Id: ", &str))
1557 return 0;
1558 for (i = 0; str[i]; i++) {
1559 if (isspace(str[i]) && str[i+1] != '$')
1560 return 1;
1561 }
1562 return 0;
1563 }
1564
1565 static void ident_drain(struct ident_filter *ident, char **output_p, size_t *osize_p)
1566 {
1567 size_t to_drain = ident->left.len;
1568
1569 if (*osize_p < to_drain)
1570 to_drain = *osize_p;
1571 if (to_drain) {
1572 memcpy(*output_p, ident->left.buf, to_drain);
1573 strbuf_remove(&ident->left, 0, to_drain);
1574 *output_p += to_drain;
1575 *osize_p -= to_drain;
1576 }
1577 if (!ident->left.len)
1578 ident->state = 0;
1579 }
1580
1581 static int ident_filter_fn(struct stream_filter *filter,
1582 const char *input, size_t *isize_p,
1583 char *output, size_t *osize_p)
1584 {
1585 struct ident_filter *ident = (struct ident_filter *)filter;
1586 static const char head[] = "$Id";
1587
1588 if (!input) {
1589 /* drain upon eof */
1590 switch (ident->state) {
1591 default:
1592 strbuf_add(&ident->left, head, ident->state);
1593 case IDENT_SKIPPING:
1594 /* fallthru */
1595 case IDENT_DRAINING:
1596 ident_drain(ident, &output, osize_p);
1597 }
1598 return 0;
1599 }
1600
1601 while (*isize_p || (ident->state == IDENT_DRAINING)) {
1602 int ch;
1603
1604 if (ident->state == IDENT_DRAINING) {
1605 ident_drain(ident, &output, osize_p);
1606 if (!*osize_p)
1607 break;
1608 continue;
1609 }
1610
1611 ch = *(input++);
1612 (*isize_p)--;
1613
1614 if (ident->state == IDENT_SKIPPING) {
1615 /*
1616 * Skipping until '$' or LF, but keeping them
1617 * in case it is a foreign ident.
1618 */
1619 strbuf_addch(&ident->left, ch);
1620 if (ch != '\n' && ch != '$')
1621 continue;
1622 if (ch == '$' && !is_foreign_ident(ident->left.buf)) {
1623 strbuf_setlen(&ident->left, sizeof(head) - 1);
1624 strbuf_addstr(&ident->left, ident->ident);
1625 }
1626 ident->state = IDENT_DRAINING;
1627 continue;
1628 }
1629
1630 if (ident->state < sizeof(head) &&
1631 head[ident->state] == ch) {
1632 ident->state++;
1633 continue;
1634 }
1635
1636 if (ident->state)
1637 strbuf_add(&ident->left, head, ident->state);
1638 if (ident->state == sizeof(head) - 1) {
1639 if (ch != ':' && ch != '$') {
1640 strbuf_addch(&ident->left, ch);
1641 ident->state = 0;
1642 continue;
1643 }
1644
1645 if (ch == ':') {
1646 strbuf_addch(&ident->left, ch);
1647 ident->state = IDENT_SKIPPING;
1648 } else {
1649 strbuf_addstr(&ident->left, ident->ident);
1650 ident->state = IDENT_DRAINING;
1651 }
1652 continue;
1653 }
1654
1655 strbuf_addch(&ident->left, ch);
1656 ident->state = IDENT_DRAINING;
1657 }
1658 return 0;
1659 }
1660
1661 static void ident_free_fn(struct stream_filter *filter)
1662 {
1663 struct ident_filter *ident = (struct ident_filter *)filter;
1664 strbuf_release(&ident->left);
1665 free(filter);
1666 }
1667
1668 static struct stream_filter_vtbl ident_vtbl = {
1669 ident_filter_fn,
1670 ident_free_fn,
1671 };
1672
1673 static struct stream_filter *ident_filter(const unsigned char *sha1)
1674 {
1675 struct ident_filter *ident = xmalloc(sizeof(*ident));
1676
1677 xsnprintf(ident->ident, sizeof(ident->ident),
1678 ": %s $", sha1_to_hex(sha1));
1679 strbuf_init(&ident->left, 0);
1680 ident->filter.vtbl = &ident_vtbl;
1681 ident->state = 0;
1682 return (struct stream_filter *)ident;
1683 }
1684
1685 /*
1686 * Return an appropriately constructed filter for the path, or NULL if
1687 * the contents cannot be filtered without reading the whole thing
1688 * in-core.
1689 *
1690 * Note that you would be crazy to set CRLF, smuge/clean or ident to a
1691 * large binary blob you would want us not to slurp into the memory!
1692 */
1693 struct stream_filter *get_stream_filter(const char *path, const unsigned char *sha1)
1694 {
1695 struct conv_attrs ca;
1696 struct stream_filter *filter = NULL;
1697
1698 convert_attrs(&ca, path);
1699 if (ca.drv && (ca.drv->process || ca.drv->smudge || ca.drv->clean))
1700 return NULL;
1701
1702 if (ca.crlf_action == CRLF_AUTO || ca.crlf_action == CRLF_AUTO_CRLF)
1703 return NULL;
1704
1705 if (ca.ident)
1706 filter = ident_filter(sha1);
1707
1708 if (output_eol(ca.crlf_action) == EOL_CRLF)
1709 filter = cascade_filter(filter, lf_to_crlf_filter());
1710 else
1711 filter = cascade_filter(filter, &null_filter_singleton);
1712
1713 return filter;
1714 }
1715
1716 void free_stream_filter(struct stream_filter *filter)
1717 {
1718 filter->vtbl->free(filter);
1719 }
1720
1721 int stream_filter(struct stream_filter *filter,
1722 const char *input, size_t *isize_p,
1723 char *output, size_t *osize_p)
1724 {
1725 return filter->vtbl->filter(filter, input, isize_p, output, osize_p);
1726 }
The Best VCS Around