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 //
  22 //  --check turns on checking that the working tree matches the
  23 //    files that are being modified, but doesn't apply the patch
  24 //  --stat does just a diffstat, and doesn't actually apply
  25 //  --show-files shows the directory changes
  26 //
  27 static int merge_patch = 1;
  28 static int check_index = 0;
  29 static int write_index = 0;
  30 static int diffstat = 0;
  31 static int check = 0;
  32 static int apply = 1;
  33 static int show_files = 0;
  34 static const char apply_usage[] = "git-apply [--stat] [--check] [--show-files] <patch>";
  35
  36 /*
  37  * For "diff-stat" like behaviour, we keep track of the biggest change
  38  * we've seen, and the longest filename. That allows us to do simple
  39  * scaling.
  40  */
  41 static int max_change, max_len;
  42
  43 /*
  44  * Various "current state", notably line numbers and what
  45  * file (and how) we're patching right now.. The "is_xxxx"
  46  * things are flags, where -1 means "don't know yet".
  47  */
  48 static int linenr = 1;
  49
  50 struct fragment {
  51         unsigned long oldpos, oldlines;
  52         unsigned long newpos, newlines;
  53         const char *patch;
  54         int size;
  55         struct fragment *next;
  56 };
  57
  58 struct patch {
  59         char *new_name, *old_name, *def_name;
  60         unsigned int old_mode, new_mode;
  61         int is_rename, is_copy, is_new, is_delete;
  62         int lines_added, lines_deleted;
  63         struct fragment *fragments;
  64         char *result;
  65         unsigned long resultsize;
  66         struct patch *next;
  67 };
  68
  69 #define CHUNKSIZE (8192)
  70 #define SLOP (16)
  71
  72 static void *read_patch_file(int fd, unsigned long *sizep)
  73 {
  74         unsigned long size = 0, alloc = CHUNKSIZE;
  75         void *buffer = xmalloc(alloc);
  76
  77         for (;;) {
  78                 int nr = alloc - size;
  79                 if (nr < 1024) {
  80                         alloc += CHUNKSIZE;
  81                         buffer = xrealloc(buffer, alloc);
  82                         nr = alloc - size;
  83                 }
  84                 nr = read(fd, buffer + size, nr);
  85                 if (!nr)
  86                         break;
  87                 if (nr < 0) {
  88                         if (errno == EAGAIN)
  89                                 continue;
  90                         die("git-apply: read returned %s", strerror(errno));
  91                 }
  92                 size += nr;
  93         }
  94         *sizep = size;
  95
  96         /*
  97          * Make sure that we have some slop in the buffer
  98          * so that we can do speculative "memcmp" etc, and
  99          * see to it that it is NUL-filled.
 100          */
 101         if (alloc < size + SLOP)
 102                 buffer = xrealloc(buffer, size + SLOP);
 103         memset(buffer + size, 0, SLOP);
 104         return buffer;
 105 }
 106
 107 static unsigned long linelen(const char *buffer, unsigned long size)
 108 {
 109         unsigned long len = 0;
 110         while (size--) {
 111                 len++;
 112                 if (*buffer++ == '\n')
 113                         break;
 114         }
 115         return len;
 116 }
 117
 118 static int is_dev_null(const char *str)
 119 {
 120         return !memcmp("/dev/null", str, 9) && isspace(str[9]);
 121 }
 122
 123 #define TERM_SPACE      1
 124 #define TERM_TAB        2
 125
 126 static int name_terminate(const char *name, int namelen, int c, int terminate)
 127 {
 128         if (c == ' ' && !(terminate & TERM_SPACE))
 129                 return 0;
 130         if (c == '\t' && !(terminate & TERM_TAB))
 131                 return 0;
 132
 133         return 1;
 134 }
 135
 136 static char * find_name(const char *line, char *def, int p_value, int terminate)
 137 {
 138         int len;
 139         const char *start = line;
 140         char *name;
 141
 142         for (;;) {
 143                 char c = *line;
 144
 145                 if (isspace(c)) {
 146                         if (c == '\n')
 147                                 break;
 148                         if (name_terminate(start, line-start, c, terminate))
 149                                 break;
 150                 }
 151                 line++;
 152                 if (c == '/' && !--p_value)
 153                         start = line;
 154         }
 155         if (!start)
 156                 return def;
 157         len = line - start;
 158         if (!len)
 159                 return def;
 160
 161         /*
 162          * Generally we prefer the shorter name, especially
 163          * if the other one is just a variation of that with
 164          * something else tacked on to the end (ie "file.orig"
 165          * or "file~").
 166          */
 167         if (def) {
 168                 int deflen = strlen(def);
 169                 if (deflen < len && !strncmp(start, def, deflen))
 170                         return def;
 171         }
 172
 173         name = xmalloc(len + 1);
 174         memcpy(name, start, len);
 175         name[len] = 0;
 176         free(def);
 177         return name;
 178 }
 179
 180 /*
 181  * Get the name etc info from the --/+++ lines of a traditional patch header
 182  *
 183  * NOTE! This hardcodes "-p1" behaviour in filename detection.
 184  *
 185  * FIXME! The end-of-filename heuristics are kind of screwy. For existing
 186  * files, we can happily check the index for a match, but for creating a
 187  * new file we should try to match whatever "patch" does. I have no idea.
 188  */
 189 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
 190 {
 191         int p_value = 1;
 192         char *name;
 193
 194         first += 4;     // skip "--- "
 195         second += 4;    // skip "+++ "
 196         if (is_dev_null(first)) {
 197                 patch->is_new = 1;
 198                 patch->is_delete = 0;
 199                 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
 200                 patch->new_name = name;
 201         } else if (is_dev_null(second)) {
 202                 patch->is_new = 0;
 203                 patch->is_delete = 1;
 204                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 205                 patch->old_name = name;
 206         } else {
 207                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 208                 name = find_name(second, name, p_value, TERM_SPACE | TERM_TAB);
 209                 patch->old_name = patch->new_name = name;
 210         }
 211         if (!name)
 212                 die("unable to find filename in patch at line %d", linenr);
 213 }
 214
 215 static int gitdiff_hdrend(const char *line, struct patch *patch)
 216 {
 217         return -1;
 218 }
 219
 220 /*
 221  * We're anal about diff header consistency, to make
 222  * sure that we don't end up having strange ambiguous
 223  * patches floating around.
 224  *
 225  * As a result, gitdiff_{old|new}name() will check
 226  * their names against any previous information, just
 227  * to make sure..
 228  */
 229 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
 230 {
 231         int len;
 232         const char *name;
 233
 234         if (!orig_name && !isnull)
 235                 return find_name(line, NULL, 1, 0);
 236
 237         name = "/dev/null";
 238         len = 9;
 239         if (orig_name) {
 240                 name = orig_name;
 241                 len = strlen(name);
 242                 if (isnull)
 243                         die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
 244         }
 245
 246         if (*name == '/')
 247                 goto absolute_path;
 248
 249         for (;;) {
 250                 char c = *line++;
 251                 if (c == '\n')
 252                         break;
 253                 if (c != '/')
 254                         continue;
 255 absolute_path:
 256                 if (memcmp(line, name, len) || line[len] != '\n')
 257                         break;
 258                 return orig_name;
 259         }
 260         die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
 261         return NULL;
 262 }
 263
 264 static int gitdiff_oldname(const char *line, struct patch *patch)
 265 {
 266         patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
 267         return 0;
 268 }
 269
 270 static int gitdiff_newname(const char *line, struct patch *patch)
 271 {
 272         patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
 273         return 0;
 274 }
 275
 276 static int gitdiff_oldmode(const char *line, struct patch *patch)
 277 {
 278         patch->old_mode = strtoul(line, NULL, 8);
 279         return 0;
 280 }
 281
 282 static int gitdiff_newmode(const char *line, struct patch *patch)
 283 {
 284         patch->new_mode = strtoul(line, NULL, 8);
 285         return 0;
 286 }
 287
 288 static int gitdiff_delete(const char *line, struct patch *patch)
 289 {
 290         patch->is_delete = 1;
 291         patch->old_name = patch->def_name;
 292         return gitdiff_oldmode(line, patch);
 293 }
 294
 295 static int gitdiff_newfile(const char *line, struct patch *patch)
 296 {
 297         patch->is_new = 1;
 298         patch->new_name = patch->def_name;
 299         return gitdiff_newmode(line, patch);
 300 }
 301
 302 static int gitdiff_copysrc(const char *line, struct patch *patch)
 303 {
 304         patch->is_copy = 1;
 305         patch->old_name = find_name(line, NULL, 0, 0);
 306         return 0;
 307 }
 308
 309 static int gitdiff_copydst(const char *line, struct patch *patch)
 310 {
 311         patch->is_copy = 1;
 312         patch->new_name = find_name(line, NULL, 0, 0);
 313         return 0;
 314 }
 315
 316 static int gitdiff_renamesrc(const char *line, struct patch *patch)
 317 {
 318         patch->is_rename = 1;
 319         patch->old_name = find_name(line, NULL, 0, 0);
 320         return 0;
 321 }
 322
 323 static int gitdiff_renamedst(const char *line, struct patch *patch)
 324 {
 325         patch->is_rename = 1;
 326         patch->new_name = find_name(line, NULL, 0, 0);
 327         return 0;
 328 }
 329
 330 static int gitdiff_similarity(const char *line, struct patch *patch)
 331 {
 332         return 0;
 333 }
 334
 335 static int gitdiff_dissimilarity(const char *line, struct patch *patch)
 336 {
 337         return 0;
 338 }
 339
 340 /*
 341  * This is normal for a diff that doesn't change anything: we'll fall through
 342  * into the next diff. Tell the parser to break out.
 343  */
 344 static int gitdiff_unrecognized(const char *line, struct patch *patch)
 345 {
 346         return -1;
 347 }
 348
 349 static char *git_header_name(char *line)
 350 {
 351         int len;
 352         char *name, *second;
 353
 354         /*
 355          * Find the first '/'
 356          */
 357         name = line;
 358         for (;;) {
 359                 char c = *name++;
 360                 if (c == '\n')
 361                         return NULL;
 362                 if (c == '/')
 363                         break;
 364         }
 365
 366         /*
 367          * We don't accept absolute paths (/dev/null) as possibly valid
 368          */
 369         if (name == line+1)
 370                 return NULL;
 371
 372         /*
 373          * Accept a name only if it shows up twice, exactly the same
 374          * form.
 375          */
 376         for (len = 0 ; ; len++) {
 377                 char c = name[len];
 378
 379                 switch (c) {
 380                 default:
 381                         continue;
 382                 case '\n':
 383                         break;
 384                 case '\t': case ' ':
 385                         second = name+len;
 386                         for (;;) {
 387                                 char c = *second++;
 388                                 if (c == '\n')
 389                                         return NULL;
 390                                 if (c == '/')
 391                                         break;
 392                         }
 393                         if (second[len] == '\n' && !memcmp(name, second, len)) {
 394                                 char *ret = xmalloc(len + 1);
 395                                 memcpy(ret, name, len);
 396                                 ret[len] = 0;
 397                                 return ret;
 398                         }
 399                 }
 400         }
 401         return NULL;
 402 }
 403
 404 /* Verify that we recognize the lines following a git header */
 405 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
 406 {
 407         unsigned long offset;
 408
 409         /* A git diff has explicit new/delete information, so we don't guess */
 410         patch->is_new = 0;
 411         patch->is_delete = 0;
 412
 413         /*
 414          * Some things may not have the old name in the
 415          * rest of the headers anywhere (pure mode changes,
 416          * or removing or adding empty files), so we get
 417          * the default name from the header.
 418          */
 419         patch->def_name = git_header_name(line + strlen("diff --git "));
 420
 421         line += len;
 422         size -= len;
 423         linenr++;
 424         for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
 425                 static const struct opentry {
 426                         const char *str;
 427                         int (*fn)(const char *, struct patch *);
 428                 } optable[] = {
 429                         { "@@ -", gitdiff_hdrend },
 430                         { "--- ", gitdiff_oldname },
 431                         { "+++ ", gitdiff_newname },
 432                         { "old mode ", gitdiff_oldmode },
 433                         { "new mode ", gitdiff_newmode },
 434                         { "deleted file mode ", gitdiff_delete },
 435                         { "new file mode ", gitdiff_newfile },
 436                         { "copy from ", gitdiff_copysrc },
 437                         { "copy to ", gitdiff_copydst },
 438                         { "rename old ", gitdiff_renamesrc },
 439                         { "rename new ", gitdiff_renamedst },
 440                         { "rename from ", gitdiff_renamesrc },
 441                         { "rename to ", gitdiff_renamedst },
 442                         { "similarity index ", gitdiff_similarity },
 443                         { "dissimilarity index ", gitdiff_dissimilarity },
 444                         { "", gitdiff_unrecognized },
 445                 };
 446                 int i;
 447
 448                 len = linelen(line, size);
 449                 if (!len || line[len-1] != '\n')
 450                         break;
 451                 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
 452                         const struct opentry *p = optable + i;
 453                         int oplen = strlen(p->str);
 454                         if (len < oplen || memcmp(p->str, line, oplen))
 455                                 continue;
 456                         if (p->fn(line + oplen, patch) < 0)
 457                                 return offset;
 458                         break;
 459                 }
 460         }
 461
 462         return offset;
 463 }
 464
 465 static int parse_num(const char *line, unsigned long *p)
 466 {
 467         char *ptr;
 468
 469         if (!isdigit(*line))
 470                 return 0;
 471         *p = strtoul(line, &ptr, 10);
 472         return ptr - line;
 473 }
 474
 475 static int parse_range(const char *line, int len, int offset, const char *expect,
 476                         unsigned long *p1, unsigned long *p2)
 477 {
 478         int digits, ex;
 479
 480         if (offset < 0 || offset >= len)
 481                 return -1;
 482         line += offset;
 483         len -= offset;
 484
 485         digits = parse_num(line, p1);
 486         if (!digits)
 487                 return -1;
 488
 489         offset += digits;
 490         line += digits;
 491         len -= digits;
 492
 493         *p2 = *p1;
 494         if (*line == ',') {
 495                 digits = parse_num(line+1, p2);
 496                 if (!digits)
 497                         return -1;
 498
 499                 offset += digits+1;
 500                 line += digits+1;
 501                 len -= digits+1;
 502         }
 503
 504         ex = strlen(expect);
 505         if (ex > len)
 506                 return -1;
 507         if (memcmp(line, expect, ex))
 508                 return -1;
 509
 510         return offset + ex;
 511 }
 512
 513 /*
 514  * Parse a unified diff fragment header of the
 515  * form "@@ -a,b +c,d @@"
 516  */
 517 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
 518 {
 519         int offset;
 520
 521         if (!len || line[len-1] != '\n')
 522                 return -1;
 523
 524         /* Figure out the number of lines in a fragment */
 525         offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
 526         offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
 527
 528         return offset;
 529 }
 530
 531 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
 532 {
 533         unsigned long offset, len;
 534
 535         patch->is_rename = patch->is_copy = 0;
 536         patch->is_new = patch->is_delete = -1;
 537         patch->old_mode = patch->new_mode = 0;
 538         patch->old_name = patch->new_name = NULL;
 539         for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
 540                 unsigned long nextlen;
 541
 542                 len = linelen(line, size);
 543                 if (!len)
 544                         break;
 545
 546                 /* Testing this early allows us to take a few shortcuts.. */
 547                 if (len < 6)
 548                         continue;
 549
 550                 /*
 551                  * Make sure we don't find any unconnected patch fragmants.
 552                  * That's a sign that we didn't find a header, and that a
 553                  * patch has become corrupted/broken up.
 554                  */
 555                 if (!memcmp("@@ -", line, 4)) {
 556                         struct fragment dummy;
 557                         if (parse_fragment_header(line, len, &dummy) < 0)
 558                                 continue;
 559                         error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
 560                 }
 561
 562                 if (size < len + 6)
 563                         break;
 564
 565                 /*
 566                  * Git patch? It might not have a real patch, just a rename
 567                  * or mode change, so we handle that specially
 568                  */
 569                 if (!memcmp("diff --git ", line, 11)) {
 570                         int git_hdr_len = parse_git_header(line, len, size, patch);
 571                         if (git_hdr_len < 0)
 572                                 continue;
 573                         if (!patch->old_name && !patch->new_name)
 574                                 die("git diff header lacks filename information (line %d)", linenr);
 575                         *hdrsize = git_hdr_len;
 576                         return offset;
 577                 }
 578
 579                 /** --- followed by +++ ? */
 580                 if (memcmp("--- ", line,  4) || memcmp("+++ ", line + len, 4))
 581                         continue;
 582
 583                 /*
 584                  * We only accept unified patches, so we want it to
 585                  * at least have "@@ -a,b +c,d @@\n", which is 14 chars
 586                  * minimum
 587                  */
 588                 nextlen = linelen(line + len, size - len);
 589                 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
 590                         continue;
 591
 592                 /* Ok, we'll consider it a patch */
 593                 parse_traditional_patch(line, line+len, patch);
 594                 *hdrsize = len + nextlen;
 595                 linenr += 2;
 596                 return offset;
 597         }
 598         return -1;
 599 }
 600
 601 /*
 602  * Parse a unified diff. Note that this really needs
 603  * to parse each fragment separately, since the only
 604  * way to know the difference between a "---" that is
 605  * part of a patch, and a "---" that starts the next
 606  * patch is to look at the line counts..
 607  */
 608 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
 609 {
 610         int added, deleted;
 611         int len = linelen(line, size), offset;
 612         unsigned long oldlines, newlines;
 613
 614         offset = parse_fragment_header(line, len, fragment);
 615         if (offset < 0)
 616                 return -1;
 617         oldlines = fragment->oldlines;
 618         newlines = fragment->newlines;
 619
 620         if (patch->is_new < 0) {
 621                 patch->is_new =  !oldlines;
 622                 if (!oldlines)
 623                         patch->old_name = NULL;
 624         }
 625         if (patch->is_delete < 0) {
 626                 patch->is_delete = !newlines;
 627                 if (!newlines)
 628                         patch->new_name = NULL;
 629         }
 630
 631         if (patch->is_new != !oldlines)
 632                 return error("new file depends on old contents");
 633         if (patch->is_delete != !newlines)
 634                 return error("deleted file still has contents");
 635
 636         /* Parse the thing.. */
 637         line += len;
 638         size -= len;
 639         linenr++;
 640         added = deleted = 0;
 641         for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
 642                 if (!oldlines && !newlines)
 643                         break;
 644                 len = linelen(line, size);
 645                 if (!len || line[len-1] != '\n')
 646                         return -1;
 647                 switch (*line) {
 648                 default:
 649                         return -1;
 650                 case ' ':
 651                         oldlines--;
 652                         newlines--;
 653                         break;
 654                 case '-':
 655                         deleted++;
 656                         oldlines--;
 657                         break;
 658                 case '+':
 659                         added++;
 660                         newlines--;
 661                         break;
 662                 /* We allow "\ No newline at end of file" */
 663                 case '\\':
 664                         if (len < 12 || memcmp(line, "\\ No newline", 12))
 665                                 return -1;
 666                         break;
 667                 }
 668         }
 669         patch->lines_added += added;
 670         patch->lines_deleted += deleted;
 671         return offset;
 672 }
 673
 674 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 675 {
 676         unsigned long offset = 0;
 677         struct fragment **fragp = &patch->fragments;
 678
 679         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 680                 struct fragment *fragment;
 681                 int len;
 682
 683                 fragment = xmalloc(sizeof(*fragment));
 684                 memset(fragment, 0, sizeof(*fragment));
 685                 len = parse_fragment(line, size, patch, fragment);
 686                 if (len <= 0)
 687                         die("corrupt patch at line %d", linenr);
 688
 689                 fragment->patch = line;
 690                 fragment->size = len;
 691
 692                 *fragp = fragment;
 693                 fragp = &fragment->next;
 694
 695                 offset += len;
 696                 line += len;
 697                 size -= len;
 698         }
 699         return offset;
 700 }
 701
 702 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 703 {
 704         int hdrsize, patchsize;
 705         int offset = find_header(buffer, size, &hdrsize, patch);
 706
 707         if (offset < 0)
 708                 return offset;
 709
 710         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 711
 712         return offset + hdrsize + patchsize;
 713 }
 714
 715 const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
 716 const char minuses[]= "----------------------------------------------------------------------";
 717
 718 static void show_stats(struct patch *patch)
 719 {
 720         char *name = patch->old_name;
 721         int len, max, add, del, total;
 722
 723         if (!name)
 724                 name = patch->new_name;
 725
 726         /*
 727          * "scale" the filename
 728          */
 729         len = strlen(name);
 730         max = max_len;
 731         if (max > 50)
 732                 max = 50;
 733         if (len > max)
 734                 name += len - max;
 735         len = max;
 736
 737         /*
 738          * scale the add/delete
 739          */
 740         max = max_change;
 741         if (max + len > 70)
 742                 max = 70 - len;
 743
 744         add = patch->lines_added;
 745         del = patch->lines_deleted;
 746         total = add + del;
 747
 748         total = (total * max + max_change / 2) / max_change;
 749         add = (add * max + max_change / 2) / max_change;
 750         del = total - add;
 751         printf(" %-*s |%5d %.*s%.*s\n",
 752                 len, name, patch->lines_added + patch->lines_deleted,
 753                 add, pluses, del, minuses);
 754 }
 755
 756 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
 757 {
 758         int fd;
 759         unsigned long got;
 760
 761         switch (st->st_mode & S_IFMT) {
 762         case S_IFLNK:
 763                 return readlink(path, buf, size);
 764         case S_IFREG:
 765                 fd = open(path, O_RDONLY);
 766                 if (fd < 0)
 767                         return error("unable to open %s", path);
 768                 got = 0;
 769                 for (;;) {
 770                         int ret = read(fd, buf + got, size - got);
 771                         if (ret < 0) {
 772                                 if (errno == EAGAIN)
 773                                         continue;
 774                                 break;
 775                         }
 776                         if (!ret)
 777                                 break;
 778                         got += ret;
 779                 }
 780                 close(fd);
 781                 return got;
 782
 783         default:
 784                 return -1;
 785         }
 786 }
 787
 788 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
 789 {
 790         int i;
 791         unsigned long start, backwards, forwards;
 792
 793         if (fragsize > size)
 794                 return -1;
 795
 796         start = 0;
 797         if (line > 1) {
 798                 unsigned long offset = 0;
 799                 i = line-1;
 800                 while (offset + fragsize <= size) {
 801                         if (buf[offset++] == '\n') {
 802                                 start = offset;
 803                                 if (!--i)
 804                                         break;
 805                         }
 806                 }
 807         }
 808
 809         /* Exact line number? */
 810         if (!memcmp(buf + start, fragment, fragsize))
 811                 return start;
 812
 813         /*
 814          * There's probably some smart way to do this, but I'll leave
 815          * that to the smart and beautiful people. I'm simple and stupid.
 816          */
 817         backwards = start;
 818         forwards = start;
 819         for (i = 0; ; i++) {
 820                 unsigned long try;
 821                 int n;
 822
 823                 /* "backward" */
 824                 if (i & 1) {
 825                         if (!backwards) {
 826                                 if (forwards + fragsize > size)
 827                                         break;
 828                                 continue;
 829                         }
 830                         do {
 831                                 --backwards;
 832                         } while (backwards && buf[backwards-1] != '\n');
 833                         try = backwards;
 834                 } else {
 835                         while (forwards + fragsize <= size) {
 836                                 if (buf[forwards++] == '\n')
 837                                         break;
 838                         }
 839                         try = forwards;
 840                 }
 841
 842                 if (try + fragsize > size)
 843                         continue;
 844                 if (memcmp(buf + try, fragment, fragsize))
 845                         continue;
 846                 n = (i >> 1)+1;
 847                 if (i & 1)
 848                         n = -n;
 849                 fprintf(stderr, "Fragment applied at offset %d\n", n);
 850                 return try;
 851         }
 852
 853         /*
 854          * We should start searching forward and backward.
 855          */
 856         return -1;
 857 }
 858
 859 struct buffer_desc {
 860         char *buffer;
 861         unsigned long size;
 862         unsigned long alloc;
 863 };
 864
 865 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
 866 {
 867         char *buf = desc->buffer;
 868         const char *patch = frag->patch;
 869         int offset, size = frag->size;
 870         char *old = xmalloc(size);
 871         char *new = xmalloc(size);
 872         int oldsize = 0, newsize = 0;
 873
 874         while (size > 0) {
 875                 int len = linelen(patch, size);
 876                 int plen;
 877
 878                 if (!len)
 879                         break;
 880
 881                 /*
 882                  * "plen" is how much of the line we should use for
 883                  * the actual patch data. Normally we just remove the
 884                  * first character on the line, but if the line is
 885                  * followed by "\ No newline", then we also remove the
 886                  * last one (which is the newline, of course).
 887                  */
 888                 plen = len-1;
 889                 if (len > size && patch[len] == '\\')
 890                         plen--;
 891                 switch (*patch) {
 892                 case ' ':
 893                 case '-':
 894                         memcpy(old + oldsize, patch + 1, plen);
 895                         oldsize += plen;
 896                         if (*patch == '-')
 897                                 break;
 898                 /* Fall-through for ' ' */
 899                 case '+':
 900                         memcpy(new + newsize, patch + 1, plen);
 901                         newsize += plen;
 902                         break;
 903                 case '@': case '\\':
 904                         /* Ignore it, we already handled it */
 905                         break;
 906                 default:
 907                         return -1;
 908                 }
 909                 patch += len;
 910                 size -= len;
 911         }
 912
 913         offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
 914         if (offset >= 0) {
 915                 int diff = newsize - oldsize;
 916                 unsigned long size = desc->size + diff;
 917                 unsigned long alloc = desc->alloc;
 918
 919                 if (size > alloc) {
 920                         alloc = size + 8192;
 921                         desc->alloc = alloc;
 922                         buf = xrealloc(buf, alloc);
 923                         desc->buffer = buf;
 924                 }
 925                 desc->size = size;
 926                 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
 927                 memcpy(buf + offset, new, newsize);
 928                 offset = 0;
 929         }
 930
 931         free(old);
 932         free(new);
 933         return offset;
 934 }
 935
 936 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
 937 {
 938         struct fragment *frag = patch->fragments;
 939
 940         while (frag) {
 941                 if (apply_one_fragment(desc, frag) < 0)
 942                         return error("patch failed: %s:%d", patch->old_name, frag->oldpos);
 943                 frag = frag->next;
 944         }
 945         return 0;
 946 }
 947
 948 static int apply_data(struct patch *patch, struct stat *st)
 949 {
 950         char *buf;
 951         unsigned long size, alloc;
 952         struct buffer_desc desc;
 953
 954         size = 0;
 955         alloc = 0;
 956         buf = NULL;
 957         if (patch->old_name) {
 958                 size = st->st_size;
 959                 alloc = size + 8192;
 960                 buf = xmalloc(alloc);
 961                 if (read_old_data(st, patch->old_name, buf, alloc) != size)
 962                         return error("read of %s failed", patch->old_name);
 963         }
 964
 965         desc.size = size;
 966         desc.alloc = alloc;
 967         desc.buffer = buf;
 968         if (apply_fragments(&desc, patch) < 0)
 969                 return -1;
 970         patch->result = desc.buffer;
 971         patch->resultsize = desc.size;
 972
 973         if (patch->is_delete && patch->resultsize)
 974                 return error("removal patch leaves file contents");
 975
 976         return 0;
 977 }
 978
 979 static int check_patch(struct patch *patch)
 980 {
 981         struct stat st;
 982         const char *old_name = patch->old_name;
 983         const char *new_name = patch->new_name;
 984
 985         if (old_name) {
 986                 int changed;
 987
 988                 if (lstat(old_name, &st) < 0)
 989                         return error("%s: %s\n", strerror(errno));
 990                 if (check_index) {
 991                         int pos = cache_name_pos(old_name, strlen(old_name));
 992                         if (pos < 0)
 993                                 return error("%s: does not exist in index", old_name);
 994                         changed = ce_match_stat(active_cache[pos], &st);
 995                         if (changed)
 996                                 return error("%s: does not match index", old_name);
 997                 }
 998                 if (patch->is_new < 0)
 999                         patch->is_new = 0;
1000                 if (!patch->old_mode)
1001                         patch->old_mode = st.st_mode;
1002                 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1003                         return error("%s: wrong type", old_name);
1004                 if (st.st_mode != patch->old_mode)
1005                         fprintf(stderr, "warning: %s has type %o, expected %o\n",
1006                                 old_name, st.st_mode, patch->old_mode);
1007         }
1008
1009         if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1010                 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1011                         return error("%s: already exists in index", new_name);
1012                 if (!lstat(new_name, &st))
1013                         return error("%s: already exists in working directory", new_name);
1014                 if (errno != ENOENT)
1015                         return error("%s: %s", new_name, strerror(errno));
1016                 if (!patch->new_mode)
1017                         patch->new_mode = S_IFREG | 0644;
1018         }
1019
1020         if (new_name && old_name) {
1021                 int same = !strcmp(old_name, new_name);
1022                 if (!patch->new_mode)
1023                         patch->new_mode = patch->old_mode;
1024                 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1025                         return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1026                                 patch->new_mode, new_name, patch->old_mode,
1027                                 same ? "" : " of ", same ? "" : old_name);
1028         }
1029
1030         if (apply_data(patch, &st) < 0)
1031                 return error("%s: patch does not apply", old_name);
1032         return 0;
1033 }
1034
1035 static int check_patch_list(struct patch *patch)
1036 {
1037         int error = 0;
1038
1039         for (;patch ; patch = patch->next)
1040                 error |= check_patch(patch);
1041         return error;
1042 }
1043
1044 static void show_file(int c, unsigned int mode, const char *name)
1045 {
1046         printf("%c %o %s\n", c, mode, name);
1047 }
1048
1049 static void show_file_list(struct patch *patch)
1050 {
1051         for (;patch ; patch = patch->next) {
1052                 if (patch->is_rename) {
1053                         show_file('-', patch->old_mode, patch->old_name);
1054                         show_file('+', patch->new_mode, patch->new_name);
1055                         continue;
1056                 }
1057                 if (patch->is_copy || patch->is_new) {
1058                         show_file('+', patch->new_mode, patch->new_name);
1059                         continue;
1060                 }
1061                 if (patch->is_delete) {
1062                         show_file('-', patch->old_mode, patch->old_name);
1063                         continue;
1064                 }
1065                 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
1066                         printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
1067                         continue;
1068                 }
1069                 printf("M %o %s\n", patch->old_mode, patch->old_name);
1070         }
1071 }
1072
1073 static void stat_patch_list(struct patch *patch)
1074 {
1075         int files, adds, dels;
1076
1077         for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1078                 files++;
1079                 adds += patch->lines_added;
1080                 dels += patch->lines_deleted;
1081                 show_stats(patch);
1082         }
1083
1084         printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1085 }
1086
1087 static void patch_stats(struct patch *patch)
1088 {
1089         int lines = patch->lines_added + patch->lines_deleted;
1090
1091         if (lines > max_change)
1092                 max_change = lines;
1093         if (patch->old_name) {
1094                 int len = strlen(patch->old_name);
1095                 if (len > max_len)
1096                         max_len = len;
1097         }
1098         if (patch->new_name) {
1099                 int len = strlen(patch->new_name);
1100                 if (len > max_len)
1101                         max_len = len;
1102         }
1103 }
1104
1105 static void remove_file(struct patch *patch)
1106 {
1107         if (write_index) {
1108                 if (remove_file_from_cache(patch->old_name) < 0)
1109                         die("unable to remove %s from index", patch->old_name);
1110         }
1111         unlink(patch->old_name);
1112 }
1113
1114 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1115 {
1116         struct stat st;
1117         struct cache_entry *ce;
1118         int namelen = strlen(path);
1119         unsigned ce_size = cache_entry_size(namelen);
1120
1121         if (!write_index)
1122                 return;
1123
1124         ce = xmalloc(ce_size);
1125         memset(ce, 0, ce_size);
1126         memcpy(ce->name, path, namelen);
1127         ce->ce_mode = create_ce_mode(mode);
1128         ce->ce_flags = htons(namelen);
1129         if (lstat(path, &st) < 0)
1130                 die("unable to stat newly created file %s", path);
1131         fill_stat_cache_info(ce, &st);
1132         if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1133                 die("unable to create backing store for newly created file %s", path);
1134         if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1135                 die("unable to add cache entry for %s", path);
1136 }
1137
1138 static void create_file(struct patch *patch)
1139 {
1140         const char *path = patch->new_name;
1141         unsigned mode = patch->new_mode;
1142         unsigned long size = patch->resultsize;
1143         char *buf = patch->result;
1144
1145         if (!mode)
1146                 mode = S_IFREG | 0644;
1147         if (S_ISREG(mode)) {
1148                 int fd;
1149                 mode = (mode & 0100) ? 0777 : 0666;
1150                 fd = open(path, O_WRONLY | O_CREAT | O_TRUNC, mode);
1151                 if (fd < 0)
1152                         die("unable to create file %s (%s)", path, strerror(errno));
1153                 if (write(fd, buf, size) != size)
1154                         die("unable to write file %s", path);
1155                 close(fd);
1156                 add_index_file(path, mode, buf, size);
1157                 return;
1158         }
1159         if (S_ISLNK(mode)) {
1160                 if (size && buf[size-1] == '\n')
1161                         size--;
1162                 buf[size] = 0;
1163                 if (symlink(buf, path) < 0)
1164                         die("unable to write symlink %s", path);
1165                 add_index_file(path, mode, buf, size);
1166                 return;
1167         }
1168         die("unable to write file mode %o", mode);
1169 }
1170
1171 static void write_out_one_result(struct patch *patch)
1172 {
1173         if (patch->is_delete > 0) {
1174                 remove_file(patch);
1175                 return;
1176         }
1177         if (patch->is_new > 0 || patch->is_copy) {
1178                 create_file(patch);
1179                 return;
1180         }
1181         /*
1182          * Rename or modification boils down to the same
1183          * thing: remove the old, write the new
1184          */
1185         remove_file(patch);
1186         create_file(patch);
1187 }
1188
1189 static void write_out_results(struct patch *list)
1190 {
1191         if (!list)
1192                 die("No changes");
1193
1194         while (list) {
1195                 write_out_one_result(list);
1196                 list = list->next;
1197         }
1198 }
1199
1200 static struct cache_file cache_file;
1201
1202 static int apply_patch(int fd)
1203 {
1204         int newfd;
1205         unsigned long offset, size;
1206         char *buffer = read_patch_file(fd, &size);
1207         struct patch *list = NULL, **listp = &list;
1208
1209         if (!buffer)
1210                 return -1;
1211         offset = 0;
1212         while (size > 0) {
1213                 struct patch *patch;
1214                 int nr;
1215
1216                 patch = xmalloc(sizeof(*patch));
1217                 memset(patch, 0, sizeof(*patch));
1218                 nr = parse_chunk(buffer + offset, size, patch);
1219                 if (nr < 0)
1220                         break;
1221                 patch_stats(patch);
1222                 *listp = patch;
1223                 listp = &patch->next;
1224                 offset += nr;
1225                 size -= nr;
1226         }
1227
1228         newfd = -1;
1229         write_index = check_index && apply;
1230         if (write_index)
1231                 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1232         if (check_index) {
1233                 if (read_cache() < 0)
1234                         die("unable to read index file");
1235         }
1236
1237         if ((check || apply) && check_patch_list(list) < 0)
1238                 exit(1);
1239
1240         if (apply)
1241                 write_out_results(list);
1242
1243         if (write_index) {
1244                 if (write_cache(newfd, active_cache, active_nr) ||
1245                     commit_index_file(&cache_file))
1246                         die("Unable to write new cachefile");
1247         }
1248
1249         if (show_files)
1250                 show_file_list(list);
1251
1252         if (diffstat)
1253                 stat_patch_list(list);
1254
1255         free(buffer);
1256         return 0;
1257 }
1258
1259 int main(int argc, char **argv)
1260 {
1261         int i;
1262         int read_stdin = 1;
1263
1264         for (i = 1; i < argc; i++) {
1265                 const char *arg = argv[i];
1266                 int fd;
1267
1268                 if (!strcmp(arg, "-")) {
1269                         apply_patch(0);
1270                         read_stdin = 0;
1271                         continue;
1272                 }
1273                 if (!strcmp(arg, "--no-merge")) {
1274                         merge_patch = 0;
1275                         continue;
1276                 }
1277                 if (!strcmp(arg, "--stat")) {
1278                         apply = 0;
1279                         diffstat = 1;
1280                         continue;
1281                 }
1282                 if (!strcmp(arg, "--check")) {
1283                         apply = 0;
1284                         check = 1;
1285                         continue;
1286                 }
1287                 if (!strcmp(arg, "--index")) {
1288                         check_index = 1;
1289                         continue;
1290                 }
1291                 if (!strcmp(arg, "--show-files")) {
1292                         show_files = 1;
1293                         continue;
1294                 }
1295                 fd = open(arg, O_RDONLY);
1296                 if (fd < 0)
1297                         usage(apply_usage);
1298                 read_stdin = 0;
1299                 apply_patch(fd);
1300                 close(fd);
1301         }
1302         if (read_stdin)
1303                 apply_patch(0);
1304         return 0;
1305 }