apply.c

   1 /*
   2  * apply.c
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  *
   6  * This applies patches on top of some (arbitrary) version of the SCM.
   7  *
   8  * NOTE! It does all its work in the index file, and only cares about
   9  * the files in the working directory if you tell it to "merge" the
  10  * patch apply.
  11  *
  12  * Even when merging it always takes the source from the index, and
  13  * uses the working tree as a "branch" for a 3-way merge.
  14  */
  15 #include <ctype.h>
  16
  17 #include "cache.h"
  18
  19 // We default to the merge behaviour, since that's what most people would
  20 // expect
  21 static int merge_patch = 1;
  22 static const char apply_usage[] = "git-apply <patch>";
  23
  24 /*
  25  * Various "current state", notably line numbers and what
  26  * file (and how) we're patching right now.. The "is_xxxx"
  27  * things are flags, where -1 means "don't know yet".
  28  */
  29 static int linenr = 1;
  30 static char *def_name = NULL;
  31
  32 struct fragment {
  33         unsigned long oldpos, oldlines;
  34         unsigned long newpos, newlines;
  35         const char *patch;
  36         int size;
  37         struct fragment *next;
  38 };
  39
  40 struct patch {
  41         char *new_name, *old_name;
  42         unsigned int old_mode, new_mode;
  43         int is_rename, is_copy, is_new, is_delete;
  44         struct fragment *fragments;
  45         struct patch *next;
  46 };
  47
  48 #define CHUNKSIZE (8192)
  49 #define SLOP (16)
  50
  51 static void *read_patch_file(int fd, unsigned long *sizep)
  52 {
  53         unsigned long size = 0, alloc = CHUNKSIZE;
  54         void *buffer = xmalloc(alloc);
  55
  56         for (;;) {
  57                 int nr = alloc - size;
  58                 if (nr < 1024) {
  59                         alloc += CHUNKSIZE;
  60                         buffer = xrealloc(buffer, alloc);
  61                         nr = alloc - size;
  62                 }
  63                 nr = read(fd, buffer + size, nr);
  64                 if (!nr)
  65                         break;
  66                 if (nr < 0) {
  67                         if (errno == EAGAIN)
  68                                 continue;
  69                         die("git-apply: read returned %s", strerror(errno));
  70                 }
  71                 size += nr;
  72         }
  73         *sizep = size;
  74
  75         /*
  76          * Make sure that we have some slop in the buffer
  77          * so that we can do speculative "memcmp" etc, and
  78          * see to it that it is NUL-filled.
  79          */
  80         if (alloc < size + SLOP)
  81                 buffer = xrealloc(buffer, size + SLOP);
  82         memset(buffer + size, 0, SLOP);
  83         return buffer;
  84 }
  85
  86 static unsigned long linelen(char *buffer, unsigned long size)
  87 {
  88         unsigned long len = 0;
  89         while (size--) {
  90                 len++;
  91                 if (*buffer++ == '\n')
  92                         break;
  93         }
  94         return len;
  95 }
  96
  97 static int is_dev_null(const char *str)
  98 {
  99         return !memcmp("/dev/null", str, 9) && isspace(str[9]);
 100 }
 101
 102 #define TERM_EXIST      1
 103 #define TERM_SPACE      2
 104 #define TERM_TAB        4
 105
 106 static int name_terminate(const char *name, int namelen, int c, int terminate)
 107 {
 108         if (c == ' ' && !(terminate & TERM_SPACE))
 109                 return 0;
 110         if (c == '\t' && !(terminate & TERM_TAB))
 111                 return 0;
 112
 113         /*
 114          * Do we want an existing name? Return false and
 115          * continue if it's not there.
 116          */
 117         if (terminate & TERM_EXIST)
 118                 return cache_name_pos(name, namelen) >= 0;
 119
 120         return 1;
 121 }
 122
 123 static char * find_name(const char *line, char *def, int p_value, int terminate)
 124 {
 125         int len;
 126         const char *start = line;
 127         char *name;
 128
 129         for (;;) {
 130                 char c = *line;
 131
 132                 if (isspace(c)) {
 133                         if (c == '\n')
 134                                 break;
 135                         if (name_terminate(start, line-start, c, terminate))
 136                                 break;
 137                 }
 138                 line++;
 139                 if (c == '/' && !--p_value)
 140                         start = line;
 141         }
 142         if (!start)
 143                 return def;
 144         len = line - start;
 145         if (!len)
 146                 return def;
 147
 148         /*
 149          * Generally we prefer the shorter name, especially
 150          * if the other one is just a variation of that with
 151          * something else tacked on to the end (ie "file.orig"
 152          * or "file~").
 153          */
 154         if (def) {
 155                 int deflen = strlen(def);
 156                 if (deflen < len && !strncmp(start, def, deflen))
 157                         return def;
 158         }
 159
 160         name = xmalloc(len + 1);
 161         memcpy(name, start, len);
 162         name[len] = 0;
 163         free(def);
 164         return name;
 165 }
 166
 167 /*
 168  * Get the name etc info from the --/+++ lines of a traditional patch header
 169  *
 170  * NOTE! This hardcodes "-p1" behaviour in filename detection.
 171  *
 172  * FIXME! The end-of-filename heuristics are kind of screwy. For existing
 173  * files, we can happily check the index for a match, but for creating a
 174  * new file we should try to match whatever "patch" does. I have no idea.
 175  */
 176 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
 177 {
 178         int p_value = 1;
 179         char *name;
 180
 181         first += 4;     // skip "--- "
 182         second += 4;    // skip "+++ "
 183         if (is_dev_null(first)) {
 184                 patch->is_new = 1;
 185                 patch->is_delete = 0;
 186                 name = find_name(second, def_name, p_value, TERM_SPACE | TERM_TAB);
 187                 patch->new_name = name;
 188         } else if (is_dev_null(second)) {
 189                 patch->is_new = 0;
 190                 patch->is_delete = 1;
 191                 name = find_name(first, def_name, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
 192                 patch->old_name = name;
 193         } else {
 194                 name = find_name(first, def_name, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
 195                 name = find_name(second, name, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
 196                 patch->old_name = patch->new_name = name;
 197         }
 198         if (!name)
 199                 die("unable to find filename in patch at line %d", linenr);
 200 }
 201
 202 static int gitdiff_hdrend(const char *line, struct patch *patch)
 203 {
 204         return -1;
 205 }
 206
 207 /*
 208  * We're anal about diff header consistency, to make
 209  * sure that we don't end up having strange ambiguous
 210  * patches floating around.
 211  *
 212  * As a result, gitdiff_{old|new}name() will check
 213  * their names against any previous information, just
 214  * to make sure..
 215  */
 216 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
 217 {
 218         int len;
 219         const char *name;
 220
 221         if (!orig_name && !isnull)
 222                 return find_name(line, NULL, 1, 0);
 223
 224         name = "/dev/null";
 225         len = 9;
 226         if (orig_name) {
 227                 name = orig_name;
 228                 len = strlen(name);
 229                 if (isnull)
 230                         die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
 231         }
 232
 233         if (*name == '/')
 234                 goto absolute_path;
 235
 236         for (;;) {
 237                 char c = *line++;
 238                 if (c == '\n')
 239                         break;
 240                 if (c != '/')
 241                         continue;
 242 absolute_path:
 243                 if (memcmp(line, name, len) || line[len] != '\n')
 244                         break;
 245                 return orig_name;
 246         }
 247         die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
 248         return NULL;
 249 }
 250
 251 static int gitdiff_oldname(const char *line, struct patch *patch)
 252 {
 253         patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
 254         return 0;
 255 }
 256
 257 static int gitdiff_newname(const char *line, struct patch *patch)
 258 {
 259         patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
 260         return 0;
 261 }
 262
 263 static int gitdiff_oldmode(const char *line, struct patch *patch)
 264 {
 265         patch->old_mode = strtoul(line, NULL, 8);
 266         return 0;
 267 }
 268
 269 static int gitdiff_newmode(const char *line, struct patch *patch)
 270 {
 271         patch->new_mode = strtoul(line, NULL, 8);
 272         return 0;
 273 }
 274
 275 static int gitdiff_delete(const char *line, struct patch *patch)
 276 {
 277         patch->is_delete = 1;
 278         return gitdiff_oldmode(line, patch);
 279 }
 280
 281 static int gitdiff_newfile(const char *line, struct patch *patch)
 282 {
 283         patch->is_new = 1;
 284         return gitdiff_newmode(line, patch);
 285 }
 286
 287 static int gitdiff_copysrc(const char *line, struct patch *patch)
 288 {
 289         patch->is_copy = 1;
 290         patch->old_name = find_name(line, NULL, 0, 0);
 291         return 0;
 292 }
 293
 294 static int gitdiff_copydst(const char *line, struct patch *patch)
 295 {
 296         patch->is_copy = 1;
 297         patch->new_name = find_name(line, NULL, 0, 0);
 298         return 0;
 299 }
 300
 301 static int gitdiff_renamesrc(const char *line, struct patch *patch)
 302 {
 303         patch->is_rename = 1;
 304         patch->old_name = find_name(line, NULL, 0, 0);
 305         return 0;
 306 }
 307
 308 static int gitdiff_renamedst(const char *line, struct patch *patch)
 309 {
 310         patch->is_rename = 1;
 311         patch->new_name = find_name(line, NULL, 0, 0);
 312         return 0;
 313 }
 314
 315 static int gitdiff_similarity(const char *line, struct patch *patch)
 316 {
 317         return 0;
 318 }
 319
 320 /*
 321  * This is normal for a diff that doesn't change anything: we'll fall through
 322  * into the next diff. Tell the parser to break out.
 323  */
 324 static int gitdiff_unrecognized(const char *line, struct patch *patch)
 325 {
 326         return -1;
 327 }
 328
 329 /* Verify that we recognize the lines following a git header */
 330 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
 331 {
 332         unsigned long offset;
 333
 334         /* A git diff has explicit new/delete information, so we don't guess */
 335         patch->is_new = 0;
 336         patch->is_delete = 0;
 337
 338         line += len;
 339         size -= len;
 340         linenr++;
 341         for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
 342                 static const struct opentry {
 343                         const char *str;
 344                         int (*fn)(const char *, struct patch *);
 345                 } optable[] = {
 346                         { "@@ -", gitdiff_hdrend },
 347                         { "--- ", gitdiff_oldname },
 348                         { "+++ ", gitdiff_newname },
 349                         { "old mode ", gitdiff_oldmode },
 350                         { "new mode ", gitdiff_newmode },
 351                         { "deleted file mode ", gitdiff_delete },
 352                         { "new file mode ", gitdiff_newfile },
 353                         { "copy from ", gitdiff_copysrc },
 354                         { "copy to ", gitdiff_copydst },
 355                         { "rename from ", gitdiff_renamesrc },
 356                         { "rename to ", gitdiff_renamedst },
 357                         { "similarity index ", gitdiff_similarity },
 358                         { "", gitdiff_unrecognized },
 359                 };
 360                 int i;
 361
 362                 len = linelen(line, size);
 363                 if (!len || line[len-1] != '\n')
 364                         break;
 365                 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
 366                         const struct opentry *p = optable + i;
 367                         int oplen = strlen(p->str);
 368                         if (len < oplen || memcmp(p->str, line, oplen))
 369                                 continue;
 370                         if (p->fn(line + oplen, patch) < 0)
 371                                 return offset;
 372                         break;
 373                 }
 374         }
 375
 376         return offset;
 377 }
 378
 379 static int parse_num(const char *line, int len, int offset, const char *expect, unsigned long *p)
 380 {
 381         char *ptr;
 382         int digits, ex;
 383
 384         if (offset < 0 || offset >= len)
 385                 return -1;
 386         line += offset;
 387         len -= offset;
 388
 389         if (!isdigit(*line))
 390                 return -1;
 391         *p = strtoul(line, &ptr, 10);
 392
 393         digits = ptr - line;
 394
 395         offset += digits;
 396         line += digits;
 397         len -= digits;
 398
 399         ex = strlen(expect);
 400         if (ex > len)
 401                 return -1;
 402         if (memcmp(line, expect, ex))
 403                 return -1;
 404
 405         return offset + ex;
 406 }
 407
 408 /*
 409  * Parse a unified diff fragment header of the
 410  * form "@@ -a,b +c,d @@"
 411  */
 412 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
 413 {
 414         int offset;
 415
 416         if (!len || line[len-1] != '\n')
 417                 return -1;
 418
 419         /* Figure out the number of lines in a fragment */
 420         offset = parse_num(line, len, 4, ",", &fragment->oldpos);
 421         offset = parse_num(line, len, offset, " +", &fragment->oldlines);
 422         offset = parse_num(line, len, offset, ",", &fragment->newpos);
 423         offset = parse_num(line, len, offset, " @@", &fragment->newlines);
 424
 425         return offset;
 426 }
 427
 428 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
 429 {
 430         unsigned long offset, len;
 431
 432         patch->is_rename = patch->is_copy = 0;
 433         patch->is_new = patch->is_delete = -1;
 434         patch->old_mode = patch->new_mode = 0;
 435         patch->old_name = patch->new_name = NULL;
 436         for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
 437                 unsigned long nextlen;
 438
 439                 len = linelen(line, size);
 440                 if (!len)
 441                         break;
 442
 443                 /* Testing this early allows us to take a few shortcuts.. */
 444                 if (len < 6)
 445                         continue;
 446
 447                 /*
 448                  * Make sure we don't find any unconnected patch fragmants.
 449                  * That's a sign that we didn't find a header, and that a
 450                  * patch has become corrupted/broken up.
 451                  */
 452                 if (!memcmp("@@ -", line, 4)) {
 453                         struct fragment dummy;
 454                         if (parse_fragment_header(line, len, &dummy) < 0)
 455                                 continue;
 456                         error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
 457                 }
 458
 459                 if (size < len + 6)
 460                         break;
 461
 462                 /*
 463                  * Git patch? It might not have a real patch, just a rename
 464                  * or mode change, so we handle that specially
 465                  */
 466                 if (!memcmp("diff --git ", line, 11)) {
 467                         int git_hdr_len = parse_git_header(line, len, size, patch);
 468                         if (git_hdr_len < 0)
 469                                 continue;
 470
 471                         *hdrsize = git_hdr_len;
 472                         return offset;
 473                 }
 474
 475                 /** --- followed by +++ ? */
 476                 if (memcmp("--- ", line,  4) || memcmp("+++ ", line + len, 4))
 477                         continue;
 478
 479                 /*
 480                  * We only accept unified patches, so we want it to
 481                  * at least have "@@ -a,b +c,d @@\n", which is 14 chars
 482                  * minimum
 483                  */
 484                 nextlen = linelen(line + len, size - len);
 485                 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
 486                         continue;
 487
 488                 /* Ok, we'll consider it a patch */
 489                 parse_traditional_patch(line, line+len, patch);
 490                 *hdrsize = len + nextlen;
 491                 linenr += 2;
 492                 return offset;
 493         }
 494         return -1;
 495 }
 496
 497 /*
 498  * Parse a unified diff. Note that this really needs
 499  * to parse each fragment separately, since the only
 500  * way to know the difference between a "---" that is
 501  * part of a patch, and a "---" that starts the next
 502  * patch is to look at the line counts..
 503  */
 504 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
 505 {
 506         int len = linelen(line, size), offset;
 507         unsigned long pos[4], oldlines, newlines;
 508
 509         offset = parse_fragment_header(line, len, fragment);
 510         if (offset < 0)
 511                 return -1;
 512         oldlines = fragment->oldlines;
 513         newlines = fragment->newlines;
 514
 515         if (patch->is_new < 0 && (pos[0] || oldlines))
 516                 patch->is_new = 0;
 517         if (patch->is_delete < 0 && (pos[1] || newlines))
 518                 patch->is_delete = 0;
 519
 520         /* Parse the thing.. */
 521         line += len;
 522         size -= len;
 523         linenr++;
 524         for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
 525                 if (!oldlines && !newlines)
 526                         break;
 527                 len = linelen(line, size);
 528                 if (!len || line[len-1] != '\n')
 529                         return -1;
 530                 switch (*line) {
 531                 default:
 532                         return -1;
 533                 case ' ':
 534                         oldlines--;
 535                         newlines--;
 536                         break;
 537                 case '-':
 538                         oldlines--;
 539                         break;
 540                 case '+':
 541                         newlines--;
 542                         break;
 543                 }
 544         }
 545         return offset;
 546 }
 547
 548 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 549 {
 550         unsigned long offset = 0;
 551         struct fragment **fragp = &patch->fragments;
 552
 553         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 554                 struct fragment *fragment;
 555                 int len;
 556
 557                 fragment = xmalloc(sizeof(*fragment));
 558                 memset(fragment, 0, sizeof(*fragment));
 559                 len = parse_fragment(line, size, patch, fragment);
 560                 if (len <= 0)
 561                         die("corrupt patch at line %d", linenr);
 562
 563                 fragment->patch = line;
 564                 fragment->size = len;
 565
 566                 *fragp = fragment;
 567                 fragp = &fragment->next;
 568
 569                 offset += len;
 570                 line += len;
 571                 size -= len;
 572         }
 573         return offset;
 574 }
 575
 576 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 577 {
 578         int hdrsize, patchsize;
 579         int offset = find_header(buffer, size, &hdrsize, patch);
 580
 581         if (offset < 0)
 582                 return offset;
 583
 584         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 585
 586         return offset + hdrsize + patchsize;
 587 }
 588
 589 static void apply_patch_list(struct patch *patch)
 590 {
 591         if (!patch)
 592                 die("no patch found");
 593         do {
 594                 const char *old_name = patch->old_name;
 595                 const char *new_name = patch->new_name;
 596                 struct fragment *frag;
 597
 598                 if (old_name) {
 599                         if (cache_name_pos(old_name, strlen(old_name)) < 0)
 600                                 die("file %s does not exist", old_name);
 601                         if (patch->is_new < 0)
 602                                 patch->is_new = 0;
 603                 }
 604                 if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
 605                         if (cache_name_pos(new_name, strlen(new_name)) >= 0)
 606                                 die("file %s already exists", new_name);
 607                 }
 608
 609                 printf("Applying patch to %s\n", new_name);
 610                 printf("  new=%d delete=%d copy=%d rename=%d\n",
 611                         patch->is_new, patch->is_delete, patch->is_copy, patch->is_rename);
 612                 if (patch->old_mode != patch->new_mode)
 613                         printf("  %o->%o\n", patch->old_mode, patch->new_mode);
 614                 frag = patch->fragments;
 615                 while (frag) {
 616                         printf("Fragment %lu,%lu -> %lu,%lu\n%.*s",
 617                                 frag->oldpos, frag->oldlines,
 618                                 frag->newpos, frag->newlines,
 619                                 frag->size, frag->patch);
 620                         frag = frag->next;
 621                 }
 622
 623         } while ((patch = patch->next) != NULL);
 624 }
 625
 626 static int apply_patch(int fd)
 627 {
 628         unsigned long offset, size;
 629         char *buffer = read_patch_file(fd, &size);
 630         struct patch *list = NULL, **listp = &list;
 631
 632         if (!buffer)
 633                 return -1;
 634         offset = 0;
 635         while (size > 0) {
 636                 struct patch *patch;
 637                 int nr;
 638
 639                 patch = xmalloc(sizeof(*patch));
 640                 memset(patch, 0, sizeof(*patch));
 641                 nr = parse_chunk(buffer + offset, size, patch);
 642                 if (nr < 0)
 643                         break;
 644                 *listp = patch;
 645                 listp = &patch->next;
 646                 offset += nr;
 647                 size -= nr;
 648         }
 649
 650         apply_patch_list(list);
 651
 652         free(buffer);
 653         return 0;
 654 }
 655
 656 int main(int argc, char **argv)
 657 {
 658         int i;
 659         int read_stdin = 1;
 660
 661         if (read_cache() < 0)
 662                 die("unable to read index file");
 663
 664         for (i = 1; i < argc; i++) {
 665                 const char *arg = argv[i];
 666                 int fd;
 667
 668                 if (!strcmp(arg, "-")) {
 669                         apply_patch(0);
 670                         read_stdin = 0;
 671                         continue;
 672                 }
 673                 if (!strcmp(arg, "--no-merge")) {
 674                         merge_patch = 0;
 675                         continue;
 676                 }
 677                 fd = open(arg, O_RDONLY);
 678                 if (fd < 0)
 679                         usage(apply_usage);
 680                 read_stdin = 0;
 681                 apply_patch(fd);
 682                 close(fd);
 683         }
 684         if (read_stdin)
 685                 apply_patch(0);
 686         return 0;
 687 }