range-diff.c

   1 #include "cache.h"
   2 #include "range-diff.h"
   3 #include "string-list.h"
   4 #include "run-command.h"
   5 #include "strvec.h"
   6 #include "hashmap.h"
   7 #include "xdiff-interface.h"
   8 #include "linear-assignment.h"
   9 #include "diffcore.h"
  10 #include "commit.h"
  11 #include "pretty.h"
  12 #include "userdiff.h"
  13 #include "apply.h"
  14
  15 struct patch_util {
  16         /* For the search for an exact match */
  17         struct hashmap_entry e;
  18         const char *diff, *patch;
  19
  20         int i, shown;
  21         int diffsize;
  22         size_t diff_offset;
  23         /* the index of the matching item in the other branch, or -1 */
  24         int matching;
  25         struct object_id oid;
  26 };
  27
  28 static size_t find_end_of_line(char *buffer, unsigned long size)
  29 {
  30         char *eol = memchr(buffer, '\n', size);
  31
  32         if (!eol)
  33                 return size;
  34
  35         *eol = '\0';
  36         return eol + 1 - buffer;
  37 }
  38
  39 /*
  40  * Reads the patches into a string list, with the `util` field being populated
  41  * as struct object_id (will need to be free()d).
  42  */
  43 static int read_patches(const char *range, struct string_list *list,
  44                         const struct strvec *other_arg)
  45 {
  46         struct child_process cp = CHILD_PROCESS_INIT;
  47         struct strbuf buf = STRBUF_INIT, contents = STRBUF_INIT;
  48         struct patch_util *util = NULL;
  49         int in_header = 1;
  50         char *line, *current_filename = NULL;
  51         int offset, len;
  52         size_t size;
  53
  54         strvec_pushl(&cp.args, "log", "--no-color", "-p", "--no-merges",
  55                      "--reverse", "--date-order", "--decorate=no",
  56                      "--no-prefix",
  57                      /*
  58                       * Choose indicators that are not used anywhere
  59                       * else in diffs, but still look reasonable
  60                       * (e.g. will not be confusing when debugging)
  61                       */
  62                      "--output-indicator-new=>",
  63                      "--output-indicator-old=<",
  64                      "--output-indicator-context=#",
  65                      "--no-abbrev-commit",
  66                      "--pretty=medium",
  67                      "--notes",
  68                      NULL);
  69         if (other_arg)
  70                 strvec_pushv(&cp.args, other_arg->v);
  71         strvec_push(&cp.args, range);
  72         cp.out = -1;
  73         cp.no_stdin = 1;
  74         cp.git_cmd = 1;
  75
  76         if (start_command(&cp))
  77                 return error_errno(_("could not start `log`"));
  78         if (strbuf_read(&contents, cp.out, 0) < 0) {
  79                 error_errno(_("could not read `log` output"));
  80                 finish_command(&cp);
  81                 return -1;
  82         }
  83
  84         line = contents.buf;
  85         size = contents.len;
  86         for (offset = 0; size > 0; offset += len, size -= len, line += len) {
  87                 const char *p;
  88
  89                 len = find_end_of_line(line, size);
  90                 line[len - 1] = '\0';
  91                 if (skip_prefix(line, "commit ", &p)) {
  92                         if (util) {
  93                                 string_list_append(list, buf.buf)->util = util;
  94                                 strbuf_reset(&buf);
  95                         }
  96                         util = xcalloc(sizeof(*util), 1);
  97                         if (get_oid(p, &util->oid)) {
  98                                 error(_("could not parse commit '%s'"), p);
  99                                 free(util);
 100                                 string_list_clear(list, 1);
 101                                 strbuf_release(&buf);
 102                                 strbuf_release(&contents);
 103                                 finish_command(&cp);
 104                                 return -1;
 105                         }
 106                         util->matching = -1;
 107                         in_header = 1;
 108                         continue;
 109                 }
 110
 111                 if (!util) {
 112                         error(_("could not parse first line of `log` output: "
 113                                 "did not start with 'commit ': '%s'"),
 114                               line);
 115                         string_list_clear(list, 1);
 116                         strbuf_release(&buf);
 117                         strbuf_release(&contents);
 118                         finish_command(&cp);
 119                         return -1;
 120                 }
 121
 122                 if (starts_with(line, "diff --git")) {
 123                         struct patch patch = { 0 };
 124                         struct strbuf root = STRBUF_INIT;
 125                         int linenr = 0;
 126                         int orig_len;
 127
 128                         in_header = 0;
 129                         strbuf_addch(&buf, '\n');
 130                         if (!util->diff_offset)
 131                                 util->diff_offset = buf.len;
 132                         line[len - 1] = '\n';
 133                         orig_len = len;
 134                         len = parse_git_diff_header(&root, &linenr, 0, line,
 135                                                     len, size, &patch);
 136                         if (len < 0)
 137                                 die(_("could not parse git header '%.*s'"),
 138                                     orig_len, line);
 139                         strbuf_addstr(&buf, " ## ");
 140                         if (patch.is_new > 0)
 141                                 strbuf_addf(&buf, "%s (new)", patch.new_name);
 142                         else if (patch.is_delete > 0)
 143                                 strbuf_addf(&buf, "%s (deleted)", patch.old_name);
 144                         else if (patch.is_rename)
 145                                 strbuf_addf(&buf, "%s => %s", patch.old_name, patch.new_name);
 146                         else
 147                                 strbuf_addstr(&buf, patch.new_name);
 148
 149                         free(current_filename);
 150                         if (patch.is_delete > 0)
 151                                 current_filename = xstrdup(patch.old_name);
 152                         else
 153                                 current_filename = xstrdup(patch.new_name);
 154
 155                         if (patch.new_mode && patch.old_mode &&
 156                             patch.old_mode != patch.new_mode)
 157                                 strbuf_addf(&buf, " (mode change %06o => %06o)",
 158                                             patch.old_mode, patch.new_mode);
 159
 160                         strbuf_addstr(&buf, " ##");
 161                 } else if (in_header) {
 162                         if (starts_with(line, "Author: ")) {
 163                                 strbuf_addstr(&buf, " ## Metadata ##\n");
 164                                 strbuf_addstr(&buf, line);
 165                                 strbuf_addstr(&buf, "\n\n");
 166                                 strbuf_addstr(&buf, " ## Commit message ##\n");
 167                         } else if (starts_with(line, "Notes") &&
 168                                    line[strlen(line) - 1] == ':') {
 169                                 strbuf_addstr(&buf, "\n\n");
 170                                 /* strip the trailing colon */
 171                                 strbuf_addf(&buf, " ## %.*s ##\n",
 172                                             (int)(strlen(line) - 1), line);
 173                         } else if (starts_with(line, "    ")) {
 174                                 p = line + len - 2;
 175                                 while (isspace(*p) && p >= line)
 176                                         p--;
 177                                 strbuf_add(&buf, line, p - line + 1);
 178                                 strbuf_addch(&buf, '\n');
 179                         }
 180                         continue;
 181                 } else if (skip_prefix(line, "@@ ", &p)) {
 182                         p = strstr(p, "@@");
 183                         strbuf_addstr(&buf, "@@");
 184                         if (current_filename && p[2])
 185                                 strbuf_addf(&buf, " %s:", current_filename);
 186                         if (p)
 187                                 strbuf_addstr(&buf, p + 2);
 188                 } else if (!line[0])
 189                         /*
 190                          * A completely blank (not ' \n', which is context)
 191                          * line is not valid in a diff.  We skip it
 192                          * silently, because this neatly handles the blank
 193                          * separator line between commits in git-log
 194                          * output.
 195                          */
 196                         continue;
 197                 else if (line[0] == '>') {
 198                         strbuf_addch(&buf, '+');
 199                         strbuf_addstr(&buf, line + 1);
 200                 } else if (line[0] == '<') {
 201                         strbuf_addch(&buf, '-');
 202                         strbuf_addstr(&buf, line + 1);
 203                 } else if (line[0] == '#') {
 204                         strbuf_addch(&buf, ' ');
 205                         strbuf_addstr(&buf, line + 1);
 206                 } else {
 207                         strbuf_addch(&buf, ' ');
 208                         strbuf_addstr(&buf, line);
 209                 }
 210
 211                 strbuf_addch(&buf, '\n');
 212                 util->diffsize++;
 213         }
 214         strbuf_release(&contents);
 215
 216         if (util)
 217                 string_list_append(list, buf.buf)->util = util;
 218         strbuf_release(&buf);
 219         free(current_filename);
 220
 221         if (finish_command(&cp))
 222                 return -1;
 223
 224         return 0;
 225 }
 226
 227 static int patch_util_cmp(const void *dummy, const struct patch_util *a,
 228                           const struct patch_util *b, const char *keydata)
 229 {
 230         return strcmp(a->diff, keydata ? keydata : b->diff);
 231 }
 232
 233 static void find_exact_matches(struct string_list *a, struct string_list *b)
 234 {
 235         struct hashmap map = HASHMAP_INIT((hashmap_cmp_fn)patch_util_cmp, NULL);
 236         int i;
 237
 238         /* First, add the patches of a to a hash map */
 239         for (i = 0; i < a->nr; i++) {
 240                 struct patch_util *util = a->items[i].util;
 241
 242                 util->i = i;
 243                 util->patch = a->items[i].string;
 244                 util->diff = util->patch + util->diff_offset;
 245                 hashmap_entry_init(&util->e, strhash(util->diff));
 246                 hashmap_add(&map, &util->e);
 247         }
 248
 249         /* Now try to find exact matches in b */
 250         for (i = 0; i < b->nr; i++) {
 251                 struct patch_util *util = b->items[i].util, *other;
 252
 253                 util->i = i;
 254                 util->patch = b->items[i].string;
 255                 util->diff = util->patch + util->diff_offset;
 256                 hashmap_entry_init(&util->e, strhash(util->diff));
 257                 other = hashmap_remove_entry(&map, util, e, NULL);
 258                 if (other) {
 259                         if (other->matching >= 0)
 260                                 BUG("already assigned!");
 261
 262                         other->matching = i;
 263                         util->matching = other->i;
 264                 }
 265         }
 266
 267         hashmap_clear(&map);
 268 }
 269
 270 static void diffsize_consume(void *data, char *line, unsigned long len)
 271 {
 272         (*(int *)data)++;
 273 }
 274
 275 static void diffsize_hunk(void *data, long ob, long on, long nb, long nn,
 276                           const char *funcline, long funclen)
 277 {
 278         diffsize_consume(data, NULL, 0);
 279 }
 280
 281 static int diffsize(const char *a, const char *b)
 282 {
 283         xpparam_t pp = { 0 };
 284         xdemitconf_t cfg = { 0 };
 285         mmfile_t mf1, mf2;
 286         int count = 0;
 287
 288         mf1.ptr = (char *)a;
 289         mf1.size = strlen(a);
 290         mf2.ptr = (char *)b;
 291         mf2.size = strlen(b);
 292
 293         cfg.ctxlen = 3;
 294         if (!xdi_diff_outf(&mf1, &mf2,
 295                            diffsize_hunk, diffsize_consume, &count,
 296                            &pp, &cfg))
 297                 return count;
 298
 299         error(_("failed to generate diff"));
 300         return COST_MAX;
 301 }
 302
 303 static void get_correspondences(struct string_list *a, struct string_list *b,
 304                                 int creation_factor)
 305 {
 306         int n = a->nr + b->nr;
 307         int *cost, c, *a2b, *b2a;
 308         int i, j;
 309
 310         ALLOC_ARRAY(cost, st_mult(n, n));
 311         ALLOC_ARRAY(a2b, n);
 312         ALLOC_ARRAY(b2a, n);
 313
 314         for (i = 0; i < a->nr; i++) {
 315                 struct patch_util *a_util = a->items[i].util;
 316
 317                 for (j = 0; j < b->nr; j++) {
 318                         struct patch_util *b_util = b->items[j].util;
 319
 320                         if (a_util->matching == j)
 321                                 c = 0;
 322                         else if (a_util->matching < 0 && b_util->matching < 0)
 323                                 c = diffsize(a_util->diff, b_util->diff);
 324                         else
 325                                 c = COST_MAX;
 326                         cost[i + n * j] = c;
 327                 }
 328
 329                 c = a_util->matching < 0 ?
 330                         a_util->diffsize * creation_factor / 100 : COST_MAX;
 331                 for (j = b->nr; j < n; j++)
 332                         cost[i + n * j] = c;
 333         }
 334
 335         for (j = 0; j < b->nr; j++) {
 336                 struct patch_util *util = b->items[j].util;
 337
 338                 c = util->matching < 0 ?
 339                         util->diffsize * creation_factor / 100 : COST_MAX;
 340                 for (i = a->nr; i < n; i++)
 341                         cost[i + n * j] = c;
 342         }
 343
 344         for (i = a->nr; i < n; i++)
 345                 for (j = b->nr; j < n; j++)
 346                         cost[i + n * j] = 0;
 347
 348         compute_assignment(n, n, cost, a2b, b2a);
 349
 350         for (i = 0; i < a->nr; i++)
 351                 if (a2b[i] >= 0 && a2b[i] < b->nr) {
 352                         struct patch_util *a_util = a->items[i].util;
 353                         struct patch_util *b_util = b->items[a2b[i]].util;
 354
 355                         a_util->matching = a2b[i];
 356                         b_util->matching = i;
 357                 }
 358
 359         free(cost);
 360         free(a2b);
 361         free(b2a);
 362 }
 363
 364 static void output_pair_header(struct diff_options *diffopt,
 365                                int patch_no_width,
 366                                struct strbuf *buf,
 367                                struct strbuf *dashes,
 368                                struct patch_util *a_util,
 369                                struct patch_util *b_util)
 370 {
 371         struct object_id *oid = a_util ? &a_util->oid : &b_util->oid;
 372         struct commit *commit;
 373         char status;
 374         const char *color_reset = diff_get_color_opt(diffopt, DIFF_RESET);
 375         const char *color_old = diff_get_color_opt(diffopt, DIFF_FILE_OLD);
 376         const char *color_new = diff_get_color_opt(diffopt, DIFF_FILE_NEW);
 377         const char *color_commit = diff_get_color_opt(diffopt, DIFF_COMMIT);
 378         const char *color;
 379
 380         if (!dashes->len)
 381                 strbuf_addchars(dashes, '-',
 382                                 strlen(find_unique_abbrev(oid,
 383                                                           DEFAULT_ABBREV)));
 384
 385         if (!b_util) {
 386                 color = color_old;
 387                 status = '<';
 388         } else if (!a_util) {
 389                 color = color_new;
 390                 status = '>';
 391         } else if (strcmp(a_util->patch, b_util->patch)) {
 392                 color = color_commit;
 393                 status = '!';
 394         } else {
 395                 color = color_commit;
 396                 status = '=';
 397         }
 398
 399         strbuf_reset(buf);
 400         strbuf_addstr(buf, status == '!' ? color_old : color);
 401         if (!a_util)
 402                 strbuf_addf(buf, "%*s:  %s ", patch_no_width, "-", dashes->buf);
 403         else
 404                 strbuf_addf(buf, "%*d:  %s ", patch_no_width, a_util->i + 1,
 405                             find_unique_abbrev(&a_util->oid, DEFAULT_ABBREV));
 406
 407         if (status == '!')
 408                 strbuf_addf(buf, "%s%s", color_reset, color);
 409         strbuf_addch(buf, status);
 410         if (status == '!')
 411                 strbuf_addf(buf, "%s%s", color_reset, color_new);
 412
 413         if (!b_util)
 414                 strbuf_addf(buf, " %*s:  %s", patch_no_width, "-", dashes->buf);
 415         else
 416                 strbuf_addf(buf, " %*d:  %s", patch_no_width, b_util->i + 1,
 417                             find_unique_abbrev(&b_util->oid, DEFAULT_ABBREV));
 418
 419         commit = lookup_commit_reference(the_repository, oid);
 420         if (commit) {
 421                 if (status == '!')
 422                         strbuf_addf(buf, "%s%s", color_reset, color);
 423
 424                 strbuf_addch(buf, ' ');
 425                 pp_commit_easy(CMIT_FMT_ONELINE, commit, buf);
 426         }
 427         strbuf_addf(buf, "%s\n", color_reset);
 428
 429         fwrite(buf->buf, buf->len, 1, diffopt->file);
 430 }
 431
 432 static struct userdiff_driver section_headers = {
 433         .funcname = { "^ ## (.*) ##$\n"
 434                       "^.?@@ (.*)$", REG_EXTENDED }
 435 };
 436
 437 static struct diff_filespec *get_filespec(const char *name, const char *p)
 438 {
 439         struct diff_filespec *spec = alloc_filespec(name);
 440
 441         fill_filespec(spec, &null_oid, 0, 0100644);
 442         spec->data = (char *)p;
 443         spec->size = strlen(p);
 444         spec->should_munmap = 0;
 445         spec->is_stdin = 1;
 446         spec->driver = &section_headers;
 447
 448         return spec;
 449 }
 450
 451 static void patch_diff(const char *a, const char *b,
 452                        struct diff_options *diffopt)
 453 {
 454         diff_queue(&diff_queued_diff,
 455                    get_filespec("a", a), get_filespec("b", b));
 456
 457         diffcore_std(diffopt);
 458         diff_flush(diffopt);
 459 }
 460
 461 static void output(struct string_list *a, struct string_list *b,
 462                    struct diff_options *diffopt)
 463 {
 464         struct strbuf buf = STRBUF_INIT, dashes = STRBUF_INIT;
 465         int patch_no_width = decimal_width(1 + (a->nr > b->nr ? a->nr : b->nr));
 466         int i = 0, j = 0;
 467
 468         /*
 469          * We assume the user is really more interested in the second argument
 470          * ("newer" version). To that end, we print the output in the order of
 471          * the RHS (the `b` parameter). To put the LHS (the `a` parameter)
 472          * commits that are no longer in the RHS into a good place, we place
 473          * them once we have shown all of their predecessors in the LHS.
 474          */
 475
 476         while (i < a->nr || j < b->nr) {
 477                 struct patch_util *a_util, *b_util;
 478                 a_util = i < a->nr ? a->items[i].util : NULL;
 479                 b_util = j < b->nr ? b->items[j].util : NULL;
 480
 481                 /* Skip all the already-shown commits from the LHS. */
 482                 while (i < a->nr && a_util->shown)
 483                         a_util = ++i < a->nr ? a->items[i].util : NULL;
 484
 485                 /* Show unmatched LHS commit whose predecessors were shown. */
 486                 if (i < a->nr && a_util->matching < 0) {
 487                         output_pair_header(diffopt, patch_no_width,
 488                                            &buf, &dashes, a_util, NULL);
 489                         i++;
 490                         continue;
 491                 }
 492
 493                 /* Show unmatched RHS commits. */
 494                 while (j < b->nr && b_util->matching < 0) {
 495                         output_pair_header(diffopt, patch_no_width,
 496                                            &buf, &dashes, NULL, b_util);
 497                         b_util = ++j < b->nr ? b->items[j].util : NULL;
 498                 }
 499
 500                 /* Show matching LHS/RHS pair. */
 501                 if (j < b->nr) {
 502                         a_util = a->items[b_util->matching].util;
 503                         output_pair_header(diffopt, patch_no_width,
 504                                            &buf, &dashes, a_util, b_util);
 505                         if (!(diffopt->output_format & DIFF_FORMAT_NO_OUTPUT))
 506                                 patch_diff(a->items[b_util->matching].string,
 507                                            b->items[j].string, diffopt);
 508                         a_util->shown = 1;
 509                         j++;
 510                 }
 511         }
 512         strbuf_release(&buf);
 513         strbuf_release(&dashes);
 514 }
 515
 516 static struct strbuf *output_prefix_cb(struct diff_options *opt, void *data)
 517 {
 518         return data;
 519 }
 520
 521 int show_range_diff(const char *range1, const char *range2,
 522                     int creation_factor, int dual_color,
 523                     const struct diff_options *diffopt,
 524                     const struct strvec *other_arg)
 525 {
 526         int res = 0;
 527
 528         struct string_list branch1 = STRING_LIST_INIT_DUP;
 529         struct string_list branch2 = STRING_LIST_INIT_DUP;
 530
 531         if (read_patches(range1, &branch1, other_arg))
 532                 res = error(_("could not parse log for '%s'"), range1);
 533         if (!res && read_patches(range2, &branch2, other_arg))
 534                 res = error(_("could not parse log for '%s'"), range2);
 535
 536         if (!res) {
 537                 struct diff_options opts;
 538                 struct strbuf indent = STRBUF_INIT;
 539
 540                 if (diffopt)
 541                         memcpy(&opts, diffopt, sizeof(opts));
 542                 else
 543                         diff_setup(&opts);
 544
 545                 if (!opts.output_format)
 546                         opts.output_format = DIFF_FORMAT_PATCH;
 547                 opts.flags.suppress_diff_headers = 1;
 548                 opts.flags.dual_color_diffed_diffs = dual_color;
 549                 opts.flags.suppress_hunk_header_line_count = 1;
 550                 opts.output_prefix = output_prefix_cb;
 551                 strbuf_addstr(&indent, "    ");
 552                 opts.output_prefix_data = &indent;
 553                 diff_setup_done(&opts);
 554
 555                 find_exact_matches(&branch1, &branch2);
 556                 get_correspondences(&branch1, &branch2, creation_factor);
 557                 output(&branch1, &branch2, &opts);
 558
 559                 strbuf_release(&indent);
 560         }
 561
 562         string_list_clear(&branch1, 1);
 563         string_list_clear(&branch2, 1);
 564
 565         return res;
 566 }