tree-diff.c

   1 /*
   2  * Helper functions for tree diff generation
   3  */
   4 #include "cache.h"
   5 #include "diff.h"
   6 #include "diffcore.h"
   7 #include "tree.h"
   8
   9 /*
  10  * Compare two tree entries, taking into account only path/S_ISDIR(mode),
  11  * but not their sha1's.
  12  *
  13  * NOTE files and directories *always* compare differently, even when having
  14  *      the same name - thanks to base_name_compare().
  15  */
  16 static int tree_entry_pathcmp(struct tree_desc *t1, struct tree_desc *t2)
  17 {
  18         unsigned mode1, mode2;
  19         const char *path1, *path2;
  20         const unsigned char *sha1, *sha2;
  21         int cmp, pathlen1, pathlen2;
  22
  23         sha1 = tree_entry_extract(t1, &path1, &mode1);
  24         sha2 = tree_entry_extract(t2, &path2, &mode2);
  25
  26         pathlen1 = tree_entry_len(&t1->entry);
  27         pathlen2 = tree_entry_len(&t2->entry);
  28
  29         cmp = base_name_compare(path1, pathlen1, mode1, path2, pathlen2, mode2);
  30         return cmp;
  31 }
  32
  33
  34 /* convert path, t1/t2 -> opt->diff_*() callbacks */
  35 static void emit_diff(struct diff_options *opt, struct strbuf *path,
  36                       struct tree_desc *t1, struct tree_desc *t2)
  37 {
  38         unsigned int mode1 = t1 ? t1->entry.mode : 0;
  39         unsigned int mode2 = t2 ? t2->entry.mode : 0;
  40
  41         if (mode1 && mode2) {
  42                 opt->change(opt, mode1, mode2, t1->entry.sha1, t2->entry.sha1,
  43                         1, 1, path->buf, 0, 0);
  44         }
  45         else {
  46                 const unsigned char *sha1;
  47                 unsigned int mode;
  48                 int addremove;
  49
  50                 if (mode2) {
  51                         addremove = '+';
  52                         sha1 = t2->entry.sha1;
  53                         mode = mode2;
  54                 } else {
  55                         addremove = '-';
  56                         sha1 = t1->entry.sha1;
  57                         mode = mode1;
  58                 }
  59
  60                 opt->add_remove(opt, addremove, mode, sha1, 1, path->buf, 0);
  61         }
  62 }
  63
  64
  65 /* new path should be added to diff
  66  *
  67  * 3 cases on how/when it should be called and behaves:
  68  *
  69  *      !t1,  t2        -> path added, parent lacks it
  70  *       t1, !t2        -> path removed from parent
  71  *       t1,  t2        -> path modified
  72  */
  73 static void show_path(struct strbuf *base, struct diff_options *opt,
  74                       struct tree_desc *t1, struct tree_desc *t2)
  75 {
  76         unsigned mode;
  77         const char *path;
  78         int pathlen;
  79         int old_baselen = base->len;
  80         int isdir, recurse = 0, emitthis = 1;
  81
  82         /* at least something has to be valid */
  83         assert(t1 || t2);
  84
  85         if (t2) {
  86                 /* path present in resulting tree */
  87                 tree_entry_extract(t2, &path, &mode);
  88                 pathlen = tree_entry_len(&t2->entry);
  89                 isdir = S_ISDIR(mode);
  90         } else {
  91                 /*
  92                  * a path was removed - take path from parent. Also take
  93                  * mode from parent, to decide on recursion.
  94                  */
  95                 tree_entry_extract(t1, &path, &mode);
  96                 pathlen = tree_entry_len(&t1->entry);
  97
  98                 isdir = S_ISDIR(mode);
  99                 mode = 0;
 100         }
 101
 102         if (DIFF_OPT_TST(opt, RECURSIVE) && isdir) {
 103                 recurse = 1;
 104                 emitthis = DIFF_OPT_TST(opt, TREE_IN_RECURSIVE);
 105         }
 106
 107         strbuf_add(base, path, pathlen);
 108
 109         if (emitthis)
 110                 emit_diff(opt, base, t1, t2);
 111
 112         if (recurse) {
 113                 strbuf_addch(base, '/');
 114                 diff_tree_sha1(t1 ? t1->entry.sha1 : NULL,
 115                                t2 ? t2->entry.sha1 : NULL, base->buf, opt);
 116         }
 117
 118         strbuf_setlen(base, old_baselen);
 119 }
 120
 121 static void skip_uninteresting(struct tree_desc *t, struct strbuf *base,
 122                                struct diff_options *opt)
 123 {
 124         enum interesting match;
 125
 126         while (t->size) {
 127                 match = tree_entry_interesting(&t->entry, base, 0, &opt->pathspec);
 128                 if (match) {
 129                         if (match == all_entries_not_interesting)
 130                                 t->size = 0;
 131                         break;
 132                 }
 133                 update_tree_entry(t);
 134         }
 135 }
 136
 137 int diff_tree(struct tree_desc *t1, struct tree_desc *t2,
 138               const char *base_str, struct diff_options *opt)
 139 {
 140         struct strbuf base;
 141         int baselen = strlen(base_str);
 142
 143         /* Enable recursion indefinitely */
 144         opt->pathspec.recursive = DIFF_OPT_TST(opt, RECURSIVE);
 145
 146         strbuf_init(&base, PATH_MAX);
 147         strbuf_add(&base, base_str, baselen);
 148
 149         for (;;) {
 150                 int cmp;
 151
 152                 if (diff_can_quit_early(opt))
 153                         break;
 154                 if (opt->pathspec.nr) {
 155                         skip_uninteresting(t1, &base, opt);
 156                         skip_uninteresting(t2, &base, opt);
 157                 }
 158                 if (!t1->size) {
 159                         if (!t2->size)
 160                                 break;
 161                         show_path(&base, opt, /*t1=*/NULL, t2);
 162                         update_tree_entry(t2);
 163                         continue;
 164                 }
 165                 if (!t2->size) {
 166                         show_path(&base, opt, t1, /*t2=*/NULL);
 167                         update_tree_entry(t1);
 168                         continue;
 169                 }
 170
 171                 cmp = tree_entry_pathcmp(t1, t2);
 172
 173                 /* t1 = t2 */
 174                 if (cmp == 0) {
 175                         if (DIFF_OPT_TST(opt, FIND_COPIES_HARDER) ||
 176                             hashcmp(t1->entry.sha1, t2->entry.sha1) ||
 177                             (t1->entry.mode != t2->entry.mode))
 178                                 show_path(&base, opt, t1, t2);
 179
 180                         update_tree_entry(t1);
 181                         update_tree_entry(t2);
 182                 }
 183
 184                 /* t1 < t2 */
 185                 else if (cmp < 0) {
 186                         show_path(&base, opt, t1, /*t2=*/NULL);
 187                         update_tree_entry(t1);
 188                 }
 189
 190                 /* t1 > t2 */
 191                 else {
 192                         show_path(&base, opt, /*t1=*/NULL, t2);
 193                         update_tree_entry(t2);
 194                 }
 195         }
 196
 197         strbuf_release(&base);
 198         return 0;
 199 }
 200
 201 /*
 202  * Does it look like the resulting diff might be due to a rename?
 203  *  - single entry
 204  *  - not a valid previous file
 205  */
 206 static inline int diff_might_be_rename(void)
 207 {
 208         return diff_queued_diff.nr == 1 &&
 209                 !DIFF_FILE_VALID(diff_queued_diff.queue[0]->one);
 210 }
 211
 212 static void try_to_follow_renames(struct tree_desc *t1, struct tree_desc *t2, const char *base, struct diff_options *opt)
 213 {
 214         struct diff_options diff_opts;
 215         struct diff_queue_struct *q = &diff_queued_diff;
 216         struct diff_filepair *choice;
 217         int i;
 218
 219         /*
 220          * follow-rename code is very specific, we need exactly one
 221          * path. Magic that matches more than one path is not
 222          * supported.
 223          */
 224         GUARD_PATHSPEC(&opt->pathspec, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
 225 #if 0
 226         /*
 227          * We should reject wildcards as well. Unfortunately we
 228          * haven't got a reliable way to detect that 'foo\*bar' in
 229          * fact has no wildcards. nowildcard_len is merely a hint for
 230          * optimization. Let it slip for now until wildmatch is taught
 231          * about dry-run mode and returns wildcard info.
 232          */
 233         if (opt->pathspec.has_wildcard)
 234                 die("BUG:%s:%d: wildcards are not supported",
 235                     __FILE__, __LINE__);
 236 #endif
 237
 238         /* Remove the file creation entry from the diff queue, and remember it */
 239         choice = q->queue[0];
 240         q->nr = 0;
 241
 242         diff_setup(&diff_opts);
 243         DIFF_OPT_SET(&diff_opts, RECURSIVE);
 244         DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
 245         diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
 246         diff_opts.single_follow = opt->pathspec.items[0].match;
 247         diff_opts.break_opt = opt->break_opt;
 248         diff_opts.rename_score = opt->rename_score;
 249         diff_setup_done(&diff_opts);
 250         diff_tree(t1, t2, base, &diff_opts);
 251         diffcore_std(&diff_opts);
 252         free_pathspec(&diff_opts.pathspec);
 253
 254         /* Go through the new set of filepairing, and see if we find a more interesting one */
 255         opt->found_follow = 0;
 256         for (i = 0; i < q->nr; i++) {
 257                 struct diff_filepair *p = q->queue[i];
 258
 259                 /*
 260                  * Found a source? Not only do we use that for the new
 261                  * diff_queued_diff, we will also use that as the path in
 262                  * the future!
 263                  */
 264                 if ((p->status == 'R' || p->status == 'C') &&
 265                     !strcmp(p->two->path, opt->pathspec.items[0].match)) {
 266                         const char *path[2];
 267
 268                         /* Switch the file-pairs around */
 269                         q->queue[i] = choice;
 270                         choice = p;
 271
 272                         /* Update the path we use from now on.. */
 273                         path[0] = p->one->path;
 274                         path[1] = NULL;
 275                         free_pathspec(&opt->pathspec);
 276                         parse_pathspec(&opt->pathspec,
 277                                        PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL,
 278                                        PATHSPEC_LITERAL_PATH, "", path);
 279
 280                         /*
 281                          * The caller expects us to return a set of vanilla
 282                          * filepairs to let a later call to diffcore_std()
 283                          * it makes to sort the renames out (among other
 284                          * things), but we already have found renames
 285                          * ourselves; signal diffcore_std() not to muck with
 286                          * rename information.
 287                          */
 288                         opt->found_follow = 1;
 289                         break;
 290                 }
 291         }
 292
 293         /*
 294          * Then, discard all the non-relevant file pairs...
 295          */
 296         for (i = 0; i < q->nr; i++) {
 297                 struct diff_filepair *p = q->queue[i];
 298                 diff_free_filepair(p);
 299         }
 300
 301         /*
 302          * .. and re-instate the one we want (which might be either the
 303          * original one, or the rename/copy we found)
 304          */
 305         q->queue[0] = choice;
 306         q->nr = 1;
 307 }
 308
 309 int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
 310 {
 311         void *tree1, *tree2;
 312         struct tree_desc t1, t2;
 313         unsigned long size1, size2;
 314         int retval;
 315
 316         tree1 = fill_tree_descriptor(&t1, old);
 317         tree2 = fill_tree_descriptor(&t2, new);
 318         size1 = t1.size;
 319         size2 = t2.size;
 320         retval = diff_tree(&t1, &t2, base, opt);
 321         if (!*base && DIFF_OPT_TST(opt, FOLLOW_RENAMES) && diff_might_be_rename()) {
 322                 init_tree_desc(&t1, tree1, size1);
 323                 init_tree_desc(&t2, tree2, size2);
 324                 try_to_follow_renames(&t1, &t2, base, opt);
 325         }
 326         free(tree1);
 327         free(tree2);
 328         return retval;
 329 }
 330
 331 int diff_root_tree_sha1(const unsigned char *new, const char *base, struct diff_options *opt)
 332 {
 333         return diff_tree_sha1(NULL, new, base, opt);
 334 }