range-diff.c

   1 #include "cache.h"
   2 #include "range-diff.h"
   3 #include "string-list.h"
   4 #include "run-command.h"
   5 #include "argv-array.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
  14 struct patch_util {
  15         /* For the search for an exact match */
  16         struct hashmap_entry e;
  17         const char *diff, *patch;
  18
  19         int i, shown;
  20         int diffsize;
  21         size_t diff_offset;
  22         /* the index of the matching item in the other branch, or -1 */
  23         int matching;
  24         struct object_id oid;
  25 };
  26
  27 /*
  28  * Reads the patches into a string list, with the `util` field being populated
  29  * as struct object_id (will need to be free()d).
  30  */
  31 static int read_patches(const char *range, struct string_list *list)
  32 {
  33         struct child_process cp = CHILD_PROCESS_INIT;
  34         FILE *in;
  35         struct strbuf buf = STRBUF_INIT, line = STRBUF_INIT;
  36         struct patch_util *util = NULL;
  37         int in_header = 1;
  38
  39         argv_array_pushl(&cp.args, "log", "--no-color", "-p", "--no-merges",
  40                         "--reverse", "--date-order", "--decorate=no",
  41                         "--no-abbrev-commit", range,
  42                         NULL);
  43         cp.out = -1;
  44         cp.no_stdin = 1;
  45         cp.git_cmd = 1;
  46
  47         if (start_command(&cp))
  48                 return error_errno(_("could not start `log`"));
  49         in = fdopen(cp.out, "r");
  50         if (!in) {
  51                 error_errno(_("could not read `log` output"));
  52                 finish_command(&cp);
  53                 return -1;
  54         }
  55
  56         while (strbuf_getline(&line, in) != EOF) {
  57                 const char *p;
  58
  59                 if (skip_prefix(line.buf, "commit ", &p)) {
  60                         if (util) {
  61                                 string_list_append(list, buf.buf)->util = util;
  62                                 strbuf_reset(&buf);
  63                         }
  64                         util = xcalloc(sizeof(*util), 1);
  65                         if (get_oid(p, &util->oid)) {
  66                                 error(_("could not parse commit '%s'"), p);
  67                                 free(util);
  68                                 string_list_clear(list, 1);
  69                                 strbuf_release(&buf);
  70                                 strbuf_release(&line);
  71                                 fclose(in);
  72                                 finish_command(&cp);
  73                                 return -1;
  74                         }
  75                         util->matching = -1;
  76                         in_header = 1;
  77                         continue;
  78                 }
  79
  80                 if (starts_with(line.buf, "diff --git")) {
  81                         in_header = 0;
  82                         strbuf_addch(&buf, '\n');
  83                         if (!util->diff_offset)
  84                                 util->diff_offset = buf.len;
  85                         strbuf_addbuf(&buf, &line);
  86                 } else if (in_header) {
  87                         if (starts_with(line.buf, "Author: ")) {
  88                                 strbuf_addbuf(&buf, &line);
  89                                 strbuf_addstr(&buf, "\n\n");
  90                         } else if (starts_with(line.buf, "    ")) {
  91                                 strbuf_rtrim(&line);
  92                                 strbuf_addbuf(&buf, &line);
  93                                 strbuf_addch(&buf, '\n');
  94                         }
  95                         continue;
  96                 } else if (starts_with(line.buf, "@@ "))
  97                         strbuf_addstr(&buf, "@@");
  98                 else if (!line.buf[0] || starts_with(line.buf, "index "))
  99                         /*
 100                          * A completely blank (not ' \n', which is context)
 101                          * line is not valid in a diff.  We skip it
 102                          * silently, because this neatly handles the blank
 103                          * separator line between commits in git-log
 104                          * output.
 105                          *
 106                          * We also want to ignore the diff's `index` lines
 107                          * because they contain exact blob hashes in which
 108                          * we are not interested.
 109                          */
 110                         continue;
 111                 else
 112                         strbuf_addbuf(&buf, &line);
 113
 114                 strbuf_addch(&buf, '\n');
 115                 util->diffsize++;
 116         }
 117         fclose(in);
 118         strbuf_release(&line);
 119
 120         if (util)
 121                 string_list_append(list, buf.buf)->util = util;
 122         strbuf_release(&buf);
 123
 124         if (finish_command(&cp))
 125                 return -1;
 126
 127         return 0;
 128 }
 129
 130 static int patch_util_cmp(const void *dummy, const struct patch_util *a,
 131                      const struct patch_util *b, const char *keydata)
 132 {
 133         return strcmp(a->diff, keydata ? keydata : b->diff);
 134 }
 135
 136 static void find_exact_matches(struct string_list *a, struct string_list *b)
 137 {
 138         struct hashmap map;
 139         int i;
 140
 141         hashmap_init(&map, (hashmap_cmp_fn)patch_util_cmp, NULL, 0);
 142
 143         /* First, add the patches of a to a hash map */
 144         for (i = 0; i < a->nr; i++) {
 145                 struct patch_util *util = a->items[i].util;
 146
 147                 util->i = i;
 148                 util->patch = a->items[i].string;
 149                 util->diff = util->patch + util->diff_offset;
 150                 hashmap_entry_init(util, strhash(util->diff));
 151                 hashmap_add(&map, util);
 152         }
 153
 154         /* Now try to find exact matches in b */
 155         for (i = 0; i < b->nr; i++) {
 156                 struct patch_util *util = b->items[i].util, *other;
 157
 158                 util->i = i;
 159                 util->patch = b->items[i].string;
 160                 util->diff = util->patch + util->diff_offset;
 161                 hashmap_entry_init(util, strhash(util->diff));
 162                 other = hashmap_remove(&map, util, NULL);
 163                 if (other) {
 164                         if (other->matching >= 0)
 165                                 BUG("already assigned!");
 166
 167                         other->matching = i;
 168                         util->matching = other->i;
 169                 }
 170         }
 171
 172         hashmap_free(&map, 0);
 173 }
 174
 175 static void diffsize_consume(void *data, char *line, unsigned long len)
 176 {
 177         (*(int *)data)++;
 178 }
 179
 180 static int diffsize(const char *a, const char *b)
 181 {
 182         xpparam_t pp = { 0 };
 183         xdemitconf_t cfg = { 0 };
 184         mmfile_t mf1, mf2;
 185         int count = 0;
 186
 187         mf1.ptr = (char *)a;
 188         mf1.size = strlen(a);
 189         mf2.ptr = (char *)b;
 190         mf2.size = strlen(b);
 191
 192         cfg.ctxlen = 3;
 193         if (!xdi_diff_outf(&mf1, &mf2, diffsize_consume, &count, &pp, &cfg))
 194                 return count;
 195
 196         error(_("failed to generate diff"));
 197         return COST_MAX;
 198 }
 199
 200 static void get_correspondences(struct string_list *a, struct string_list *b,
 201                                 int creation_factor)
 202 {
 203         int n = a->nr + b->nr;
 204         int *cost, c, *a2b, *b2a;
 205         int i, j;
 206
 207         ALLOC_ARRAY(cost, st_mult(n, n));
 208         ALLOC_ARRAY(a2b, n);
 209         ALLOC_ARRAY(b2a, n);
 210
 211         for (i = 0; i < a->nr; i++) {
 212                 struct patch_util *a_util = a->items[i].util;
 213
 214                 for (j = 0; j < b->nr; j++) {
 215                         struct patch_util *b_util = b->items[j].util;
 216
 217                         if (a_util->matching == j)
 218                                 c = 0;
 219                         else if (a_util->matching < 0 && b_util->matching < 0)
 220                                 c = diffsize(a_util->diff, b_util->diff);
 221                         else
 222                                 c = COST_MAX;
 223                         cost[i + n * j] = c;
 224                 }
 225
 226                 c = a_util->matching < 0 ?
 227                         a_util->diffsize * creation_factor / 100 : COST_MAX;
 228                 for (j = b->nr; j < n; j++)
 229                         cost[i + n * j] = c;
 230         }
 231
 232         for (j = 0; j < b->nr; j++) {
 233                 struct patch_util *util = b->items[j].util;
 234
 235                 c = util->matching < 0 ?
 236                         util->diffsize * creation_factor / 100 : COST_MAX;
 237                 for (i = a->nr; i < n; i++)
 238                         cost[i + n * j] = c;
 239         }
 240
 241         for (i = a->nr; i < n; i++)
 242                 for (j = b->nr; j < n; j++)
 243                         cost[i + n * j] = 0;
 244
 245         compute_assignment(n, n, cost, a2b, b2a);
 246
 247         for (i = 0; i < a->nr; i++)
 248                 if (a2b[i] >= 0 && a2b[i] < b->nr) {
 249                         struct patch_util *a_util = a->items[i].util;
 250                         struct patch_util *b_util = b->items[a2b[i]].util;
 251
 252                         a_util->matching = a2b[i];
 253                         b_util->matching = i;
 254                 }
 255
 256         free(cost);
 257         free(a2b);
 258         free(b2a);
 259 }
 260
 261 static void output_pair_header(struct diff_options *diffopt,
 262                                struct strbuf *buf,
 263                                struct strbuf *dashes,
 264                                struct patch_util *a_util,
 265                                struct patch_util *b_util)
 266 {
 267         struct object_id *oid = a_util ? &a_util->oid : &b_util->oid;
 268         struct commit *commit;
 269         char status;
 270         const char *color_reset = diff_get_color_opt(diffopt, DIFF_RESET);
 271         const char *color_old = diff_get_color_opt(diffopt, DIFF_FILE_OLD);
 272         const char *color_new = diff_get_color_opt(diffopt, DIFF_FILE_NEW);
 273         const char *color_commit = diff_get_color_opt(diffopt, DIFF_COMMIT);
 274         const char *color;
 275
 276         if (!dashes->len)
 277                 strbuf_addchars(dashes, '-',
 278                                 strlen(find_unique_abbrev(oid,
 279                                                           DEFAULT_ABBREV)));
 280
 281         if (!b_util) {
 282                 color = color_old;
 283                 status = '<';
 284         } else if (!a_util) {
 285                 color = color_new;
 286                 status = '>';
 287         } else if (strcmp(a_util->patch, b_util->patch)) {
 288                 color = color_commit;
 289                 status = '!';
 290         } else {
 291                 color = color_commit;
 292                 status = '=';
 293         }
 294
 295         strbuf_reset(buf);
 296         strbuf_addstr(buf, status == '!' ? color_old : color);
 297         if (!a_util)
 298                 strbuf_addf(buf, "-:  %s ", dashes->buf);
 299         else
 300                 strbuf_addf(buf, "%d:  %s ", a_util->i + 1,
 301                             find_unique_abbrev(&a_util->oid, DEFAULT_ABBREV));
 302
 303         if (status == '!')
 304                 strbuf_addf(buf, "%s%s", color_reset, color);
 305         strbuf_addch(buf, status);
 306         if (status == '!')
 307                 strbuf_addf(buf, "%s%s", color_reset, color_new);
 308
 309         if (!b_util)
 310                 strbuf_addf(buf, " -:  %s", dashes->buf);
 311         else
 312                 strbuf_addf(buf, " %d:  %s", b_util->i + 1,
 313                             find_unique_abbrev(&b_util->oid, DEFAULT_ABBREV));
 314
 315         commit = lookup_commit_reference(the_repository, oid);
 316         if (commit) {
 317                 if (status == '!')
 318                         strbuf_addf(buf, "%s%s", color_reset, color);
 319
 320                 strbuf_addch(buf, ' ');
 321                 pp_commit_easy(CMIT_FMT_ONELINE, commit, buf);
 322         }
 323         strbuf_addf(buf, "%s\n", color_reset);
 324
 325         fwrite(buf->buf, buf->len, 1, stdout);
 326 }
 327
 328 static struct userdiff_driver no_func_name = {
 329         .funcname = { "$^", 0 }
 330 };
 331
 332 static struct diff_filespec *get_filespec(const char *name, const char *p)
 333 {
 334         struct diff_filespec *spec = alloc_filespec(name);
 335
 336         fill_filespec(spec, &null_oid, 0, 0644);
 337         spec->data = (char *)p;
 338         spec->size = strlen(p);
 339         spec->should_munmap = 0;
 340         spec->is_stdin = 1;
 341         spec->driver = &no_func_name;
 342
 343         return spec;
 344 }
 345
 346 static void patch_diff(const char *a, const char *b,
 347                               struct diff_options *diffopt)
 348 {
 349         diff_queue(&diff_queued_diff,
 350                    get_filespec("a", a), get_filespec("b", b));
 351
 352         diffcore_std(diffopt);
 353         diff_flush(diffopt);
 354 }
 355
 356 static void output(struct string_list *a, struct string_list *b,
 357                    struct diff_options *diffopt)
 358 {
 359         struct strbuf buf = STRBUF_INIT, dashes = STRBUF_INIT;
 360         int i = 0, j = 0;
 361
 362         /*
 363          * We assume the user is really more interested in the second argument
 364          * ("newer" version). To that end, we print the output in the order of
 365          * the RHS (the `b` parameter). To put the LHS (the `a` parameter)
 366          * commits that are no longer in the RHS into a good place, we place
 367          * them once we have shown all of their predecessors in the LHS.
 368          */
 369
 370         while (i < a->nr || j < b->nr) {
 371                 struct patch_util *a_util, *b_util;
 372                 a_util = i < a->nr ? a->items[i].util : NULL;
 373                 b_util = j < b->nr ? b->items[j].util : NULL;
 374
 375                 /* Skip all the already-shown commits from the LHS. */
 376                 while (i < a->nr && a_util->shown)
 377                         a_util = ++i < a->nr ? a->items[i].util : NULL;
 378
 379                 /* Show unmatched LHS commit whose predecessors were shown. */
 380                 if (i < a->nr && a_util->matching < 0) {
 381                         output_pair_header(diffopt,
 382                                            &buf, &dashes, a_util, NULL);
 383                         i++;
 384                         continue;
 385                 }
 386
 387                 /* Show unmatched RHS commits. */
 388                 while (j < b->nr && b_util->matching < 0) {
 389                         output_pair_header(diffopt,
 390                                            &buf, &dashes, NULL, b_util);
 391                         b_util = ++j < b->nr ? b->items[j].util : NULL;
 392                 }
 393
 394                 /* Show matching LHS/RHS pair. */
 395                 if (j < b->nr) {
 396                         a_util = a->items[b_util->matching].util;
 397                         output_pair_header(diffopt,
 398                                            &buf, &dashes, a_util, b_util);
 399                         if (!(diffopt->output_format & DIFF_FORMAT_NO_OUTPUT))
 400                                 patch_diff(a->items[b_util->matching].string,
 401                                            b->items[j].string, diffopt);
 402                         a_util->shown = 1;
 403                         j++;
 404                 }
 405         }
 406         strbuf_release(&buf);
 407         strbuf_release(&dashes);
 408 }
 409
 410 int show_range_diff(const char *range1, const char *range2,
 411                     int creation_factor, struct diff_options *diffopt)
 412 {
 413         int res = 0;
 414
 415         struct string_list branch1 = STRING_LIST_INIT_DUP;
 416         struct string_list branch2 = STRING_LIST_INIT_DUP;
 417
 418         if (read_patches(range1, &branch1))
 419                 res = error(_("could not parse log for '%s'"), range1);
 420         if (!res && read_patches(range2, &branch2))
 421                 res = error(_("could not parse log for '%s'"), range2);
 422
 423         if (!res) {
 424                 find_exact_matches(&branch1, &branch2);
 425                 get_correspondences(&branch1, &branch2, creation_factor);
 426                 output(&branch1, &branch2, diffopt);
 427         }
 428
 429         string_list_clear(&branch1, 1);
 430         string_list_clear(&branch2, 1);
 431
 432         return res;
 433 }