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 static void show_path(struct strbuf *base, struct diff_options *opt,
  10                       struct tree_desc *t1, struct tree_desc *t2);
  11
  12 static int compare_tree_entry(struct tree_desc *t1, struct tree_desc *t2,
  13                               struct strbuf *base, struct diff_options *opt)
  14 {
  15         unsigned mode1, mode2;
  16         const char *path1, *path2;
  17         const unsigned char *sha1, *sha2;
  18         int cmp, pathlen1, pathlen2;
  19
  20         sha1 = tree_entry_extract(t1, &path1, &mode1);
  21         sha2 = tree_entry_extract(t2, &path2, &mode2);
  22
  23         pathlen1 = tree_entry_len(&t1->entry);
  24         pathlen2 = tree_entry_len(&t2->entry);
  25
  26         /*
  27          * NOTE files and directories *always* compare differently,
  28          * even when having the same name.
  29          */
  30         cmp = base_name_compare(path1, pathlen1, mode1, path2, pathlen2, mode2);
  31         if (cmp < 0) {
  32                 show_path(base, opt, t1, /*t2=*/NULL);
  33                 return -1;
  34         }
  35         if (cmp > 0) {
  36                 show_path(base, opt, /*t1=*/NULL, t2);
  37                 return 1;
  38         }
  39         if (!DIFF_OPT_TST(opt, FIND_COPIES_HARDER) && !hashcmp(sha1, sha2) && mode1 == mode2)
  40                 return 0;
  41
  42         show_path(base, opt, t1, t2);
  43         return 0;
  44 }
  45
  46
  47 /* convert path, t1/t2 -> opt->diff_*() callbacks */
  48 static void emit_diff(struct diff_options *opt, struct strbuf *path,
  49                       struct tree_desc *t1, struct tree_desc *t2)
  50 {
  51         unsigned int mode1 = t1 ? t1->entry.mode : 0;
  52         unsigned int mode2 = t2 ? t2->entry.mode : 0;
  53
  54         if (mode1 && mode2) {
  55                 opt->change(opt, mode1, mode2, t1->entry.sha1, t2->entry.sha1,
  56                         1, 1, path->buf, 0, 0);
  57         }
  58         else {
  59                 const unsigned char *sha1;
  60                 unsigned int mode;
  61                 int addremove;
  62
  63                 if (mode2) {
  64                         addremove = '+';
  65                         sha1 = t2->entry.sha1;
  66                         mode = mode2;
  67                 } else {
  68                         addremove = '-';
  69                         sha1 = t1->entry.sha1;
  70                         mode = mode1;
  71                 }
  72
  73                 opt->add_remove(opt, addremove, mode, sha1, 1, path->buf, 0);
  74         }
  75 }
  76
  77
  78 /* new path should be added to diff
  79  *
  80  * 3 cases on how/when it should be called and behaves:
  81  *
  82  *      !t1,  t2        -> path added, parent lacks it
  83  *       t1, !t2        -> path removed from parent
  84  *       t1,  t2        -> path modified
  85  */
  86 static void show_path(struct strbuf *base, struct diff_options *opt,
  87                       struct tree_desc *t1, struct tree_desc *t2)
  88 {
  89         unsigned mode;
  90         const char *path;
  91         int pathlen;
  92         int old_baselen = base->len;
  93         int isdir, recurse = 0, emitthis = 1;
  94
  95         /* at least something has to be valid */
  96         assert(t1 || t2);
  97
  98         if (t2) {
  99                 /* path present in resulting tree */
 100                 tree_entry_extract(t2, &path, &mode);
 101                 pathlen = tree_entry_len(&t2->entry);
 102                 isdir = S_ISDIR(mode);
 103         } else {
 104                 /*
 105                  * a path was removed - take path from parent. Also take
 106                  * mode from parent, to decide on recursion.
 107                  */
 108                 tree_entry_extract(t1, &path, &mode);
 109                 pathlen = tree_entry_len(&t1->entry);
 110
 111                 isdir = S_ISDIR(mode);
 112                 mode = 0;
 113         }
 114
 115         if (DIFF_OPT_TST(opt, RECURSIVE) && isdir) {
 116                 recurse = 1;
 117                 emitthis = DIFF_OPT_TST(opt, TREE_IN_RECURSIVE);
 118         }
 119
 120         strbuf_add(base, path, pathlen);
 121
 122         if (emitthis)
 123                 emit_diff(opt, base, t1, t2);
 124
 125         if (recurse) {
 126                 strbuf_addch(base, '/');
 127                 diff_tree_sha1(t1 ? t1->entry.sha1 : NULL,
 128                                t2 ? t2->entry.sha1 : NULL, base->buf, opt);
 129         }
 130
 131         strbuf_setlen(base, old_baselen);
 132 }
 133
 134 static void skip_uninteresting(struct tree_desc *t, struct strbuf *base,
 135                                struct diff_options *opt)
 136 {
 137         enum interesting match;
 138
 139         while (t->size) {
 140                 match = tree_entry_interesting(&t->entry, base, 0, &opt->pathspec);
 141                 if (match) {
 142                         if (match == all_entries_not_interesting)
 143                                 t->size = 0;
 144                         break;
 145                 }
 146                 update_tree_entry(t);
 147         }
 148 }
 149
 150 int diff_tree(struct tree_desc *t1, struct tree_desc *t2,
 151               const char *base_str, struct diff_options *opt)
 152 {
 153         struct strbuf base;
 154         int baselen = strlen(base_str);
 155
 156         /* Enable recursion indefinitely */
 157         opt->pathspec.recursive = DIFF_OPT_TST(opt, RECURSIVE);
 158
 159         strbuf_init(&base, PATH_MAX);
 160         strbuf_add(&base, base_str, baselen);
 161
 162         for (;;) {
 163                 if (diff_can_quit_early(opt))
 164                         break;
 165                 if (opt->pathspec.nr) {
 166                         skip_uninteresting(t1, &base, opt);
 167                         skip_uninteresting(t2, &base, opt);
 168                 }
 169                 if (!t1->size) {
 170                         if (!t2->size)
 171                                 break;
 172                         show_path(&base, opt, /*t1=*/NULL, t2);
 173                         update_tree_entry(t2);
 174                         continue;
 175                 }
 176                 if (!t2->size) {
 177                         show_path(&base, opt, t1, /*t2=*/NULL);
 178                         update_tree_entry(t1);
 179                         continue;
 180                 }
 181
 182                 cmp = compare_tree_entry(t1, t2, &base, opt);
 183
 184                 /* t1 = t2 */
 185                 if (cmp == 0) {
 186                         update_tree_entry(t1);
 187                         update_tree_entry(t2);
 188                 }
 189
 190                 /* t1 < t2 */
 191                 else if (cmp < 0) {
 192                         update_tree_entry(t1);
 193                 }
 194
 195                 /* t1 > t2 */
 196                 else {
 197                         update_tree_entry(t2);
 198                 }
 199         }
 200
 201         strbuf_release(&base);
 202         return 0;
 203 }
 204
 205 /*
 206  * Does it look like the resulting diff might be due to a rename?
 207  *  - single entry
 208  *  - not a valid previous file
 209  */
 210 static inline int diff_might_be_rename(void)
 211 {
 212         return diff_queued_diff.nr == 1 &&
 213                 !DIFF_FILE_VALID(diff_queued_diff.queue[0]->one);
 214 }
 215
 216 static void try_to_follow_renames(struct tree_desc *t1, struct tree_desc *t2, const char *base, struct diff_options *opt)
 217 {
 218         struct diff_options diff_opts;
 219         struct diff_queue_struct *q = &diff_queued_diff;
 220         struct diff_filepair *choice;
 221         int i;
 222
 223         /*
 224          * follow-rename code is very specific, we need exactly one
 225          * path. Magic that matches more than one path is not
 226          * supported.
 227          */
 228         GUARD_PATHSPEC(&opt->pathspec, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
 229 #if 0
 230         /*
 231          * We should reject wildcards as well. Unfortunately we
 232          * haven't got a reliable way to detect that 'foo\*bar' in
 233          * fact has no wildcards. nowildcard_len is merely a hint for
 234          * optimization. Let it slip for now until wildmatch is taught
 235          * about dry-run mode and returns wildcard info.
 236          */
 237         if (opt->pathspec.has_wildcard)
 238                 die("BUG:%s:%d: wildcards are not supported",
 239                     __FILE__, __LINE__);
 240 #endif
 241
 242         /* Remove the file creation entry from the diff queue, and remember it */
 243         choice = q->queue[0];
 244         q->nr = 0;
 245
 246         diff_setup(&diff_opts);
 247         DIFF_OPT_SET(&diff_opts, RECURSIVE);
 248         DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
 249         diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
 250         diff_opts.single_follow = opt->pathspec.items[0].match;
 251         diff_opts.break_opt = opt->break_opt;
 252         diff_opts.rename_score = opt->rename_score;
 253         diff_setup_done(&diff_opts);
 254         diff_tree(t1, t2, base, &diff_opts);
 255         diffcore_std(&diff_opts);
 256         free_pathspec(&diff_opts.pathspec);
 257
 258         /* Go through the new set of filepairing, and see if we find a more interesting one */
 259         opt->found_follow = 0;
 260         for (i = 0; i < q->nr; i++) {
 261                 struct diff_filepair *p = q->queue[i];
 262
 263                 /*
 264                  * Found a source? Not only do we use that for the new
 265                  * diff_queued_diff, we will also use that as the path in
 266                  * the future!
 267                  */
 268                 if ((p->status == 'R' || p->status == 'C') &&
 269                     !strcmp(p->two->path, opt->pathspec.items[0].match)) {
 270                         const char *path[2];
 271
 272                         /* Switch the file-pairs around */
 273                         q->queue[i] = choice;
 274                         choice = p;
 275
 276                         /* Update the path we use from now on.. */
 277                         path[0] = p->one->path;
 278                         path[1] = NULL;
 279                         free_pathspec(&opt->pathspec);
 280                         parse_pathspec(&opt->pathspec,
 281                                        PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL,
 282                                        PATHSPEC_LITERAL_PATH, "", path);
 283
 284                         /*
 285                          * The caller expects us to return a set of vanilla
 286                          * filepairs to let a later call to diffcore_std()
 287                          * it makes to sort the renames out (among other
 288                          * things), but we already have found renames
 289                          * ourselves; signal diffcore_std() not to muck with
 290                          * rename information.
 291                          */
 292                         opt->found_follow = 1;
 293                         break;
 294                 }
 295         }
 296
 297         /*
 298          * Then, discard all the non-relevant file pairs...
 299          */
 300         for (i = 0; i < q->nr; i++) {
 301                 struct diff_filepair *p = q->queue[i];
 302                 diff_free_filepair(p);
 303         }
 304
 305         /*
 306          * .. and re-instate the one we want (which might be either the
 307          * original one, or the rename/copy we found)
 308          */
 309         q->queue[0] = choice;
 310         q->nr = 1;
 311 }
 312
 313 int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
 314 {
 315         void *tree1, *tree2;
 316         struct tree_desc t1, t2;
 317         unsigned long size1, size2;
 318         int retval;
 319
 320         tree1 = fill_tree_descriptor(&t1, old);
 321         tree2 = fill_tree_descriptor(&t2, new);
 322         size1 = t1.size;
 323         size2 = t2.size;
 324         retval = diff_tree(&t1, &t2, base, opt);
 325         if (!*base && DIFF_OPT_TST(opt, FOLLOW_RENAMES) && diff_might_be_rename()) {
 326                 init_tree_desc(&t1, tree1, size1);
 327                 init_tree_desc(&t2, tree2, size2);
 328                 try_to_follow_renames(&t1, &t2, base, opt);
 329         }
 330         free(tree1);
 331         free(tree2);
 332         return retval;
 333 }
 334
 335 int diff_root_tree_sha1(const unsigned char *new, const char *base, struct diff_options *opt)
 336 {
 337         return diff_tree_sha1(NULL, new, base, opt);
 338 }