diff.c

   1 /*
   2  * Copyright (C) 2005 Junio C Hamano
   3  */
   4 #include <sys/types.h>
   5 #include <sys/wait.h>
   6 #include <signal.h>
   7 #include <limits.h>
   8 #include "cache.h"
   9 #include "diff.h"
  10
  11 static const char *diff_opts = "-pu";
  12
  13 static const char *external_diff(void)
  14 {
  15         static const char *external_diff_cmd = NULL;
  16         static int done_preparing = 0;
  17
  18         if (done_preparing)
  19                 return external_diff_cmd;
  20
  21         /*
  22          * Default values above are meant to match the
  23          * Linux kernel development style.  Examples of
  24          * alternative styles you can specify via environment
  25          * variables are:
  26          *
  27          * GIT_DIFF_OPTS="-c";
  28          */
  29         if (gitenv("GIT_EXTERNAL_DIFF"))
  30                 external_diff_cmd = gitenv("GIT_EXTERNAL_DIFF");
  31
  32         /* In case external diff fails... */
  33         diff_opts = gitenv("GIT_DIFF_OPTS") ? : diff_opts;
  34
  35         done_preparing = 1;
  36         return external_diff_cmd;
  37 }
  38
  39 /* Help to copy the thing properly quoted for the shell safety.
  40  * any single quote is replaced with '\'', and the caller is
  41  * expected to enclose the result within a single quote pair.
  42  *
  43  * E.g.
  44  *  original     sq_expand     result
  45  *  name     ==> name      ==> 'name'
  46  *  a b      ==> a b       ==> 'a b'
  47  *  a'b      ==> a'\''b    ==> 'a'\''b'
  48  */
  49 static char *sq_expand(const char *src)
  50 {
  51         static char *buf = NULL;
  52         int cnt, c;
  53         const char *cp;
  54         char *bp;
  55
  56         /* count bytes needed to store the quoted string. */
  57         for (cnt = 1, cp = src; *cp; cnt++, cp++)
  58                 if (*cp == '\'')
  59                         cnt += 3;
  60
  61         buf = xmalloc(cnt);
  62         bp = buf;
  63         while ((c = *src++)) {
  64                 if (c != '\'')
  65                         *bp++ = c;
  66                 else {
  67                         bp = strcpy(bp, "'\\''");
  68                         bp += 4;
  69                 }
  70         }
  71         *bp = 0;
  72         return buf;
  73 }
  74
  75 static struct diff_tempfile {
  76         const char *name;
  77         char hex[41];
  78         char mode[10];
  79         char tmp_path[50];
  80 } diff_temp[2];
  81
  82 static void builtin_diff(const char *name,
  83                          struct diff_tempfile *temp)
  84 {
  85         int i, next_at;
  86         const char *git_prefix = "# mode: ";
  87         const char *diff_cmd = "diff -L'%s%s' -L'%s%s'";
  88         const char *diff_arg  = "'%s' '%s'||:"; /* "||:" is to return 0 */
  89         const char *input_name_sq[2];
  90         const char *path0[2];
  91         const char *path1[2];
  92         const char *name_sq = sq_expand(name);
  93         char *cmd;
  94
  95         /* diff_cmd and diff_arg have 6 %s in total which makes
  96          * the sum of these strings 12 bytes larger than required.
  97          * we use 2 spaces around diff-opts, and we need to count
  98          * terminating NUL, so we subtract 9 here.
  99          */
 100         int cmd_size = (strlen(diff_cmd) + strlen(diff_opts) +
 101                         strlen(diff_arg) - 9);
 102         for (i = 0; i < 2; i++) {
 103                 input_name_sq[i] = sq_expand(temp[i].name);
 104                 if (!strcmp(temp[i].name, "/dev/null")) {
 105                         path0[i] = "/dev/null";
 106                         path1[i] = "";
 107                 } else {
 108                         path0[i] = i ? "b/" : "a/";
 109                         path1[i] = name_sq;
 110                 }
 111                 cmd_size += (strlen(path0[i]) + strlen(path1[i]) +
 112                              strlen(input_name_sq[i]));
 113         }
 114
 115         cmd = xmalloc(cmd_size);
 116
 117         next_at = 0;
 118         next_at += snprintf(cmd+next_at, cmd_size-next_at,
 119                             diff_cmd,
 120                             path0[0], path1[0], path0[1], path1[1]);
 121         next_at += snprintf(cmd+next_at, cmd_size-next_at,
 122                             " %s ", diff_opts);
 123         next_at += snprintf(cmd+next_at, cmd_size-next_at,
 124                             diff_arg, input_name_sq[0], input_name_sq[1]);
 125
 126         if (!path1[0][0])
 127                 printf("%s. %s %s\n", git_prefix, temp[1].mode, name);
 128         else if (!path1[1][0])
 129                 printf("%s%s . %s\n", git_prefix, temp[0].mode, name);
 130         else {
 131                 if (strcmp(temp[0].mode, temp[1].mode))
 132                         printf("%s%s %s %s\n", git_prefix,
 133                                temp[0].mode, temp[1].mode, name);
 134
 135                 if (strncmp(temp[0].mode, temp[1].mode, 3))
 136                         /* we do not run diff between different kind
 137                          * of objects.
 138                          */
 139                         exit(0);
 140         }
 141         fflush(NULL);
 142         execlp("/bin/sh","sh", "-c", cmd, NULL);
 143 }
 144
 145 /*
 146  * Given a name and sha1 pair, if the dircache tells us the file in
 147  * the work tree has that object contents, return true, so that
 148  * prepare_temp_file() does not have to inflate and extract.
 149  */
 150 static int work_tree_matches(const char *name, const unsigned char *sha1)
 151 {
 152         struct cache_entry *ce;
 153         struct stat st;
 154         int pos, len;
 155
 156         /* We do not read the cache ourselves here, because the
 157          * benchmark with my previous version that always reads cache
 158          * shows that it makes things worse for diff-tree comparing
 159          * two linux-2.6 kernel trees in an already checked out work
 160          * tree.  This is because most diff-tree comparison deals with
 161          * only a small number of files, while reading the cache is
 162          * expensive for a large project, and its cost outweighs the
 163          * savings we get by not inflating the object to a temporary
 164          * file.  Practically, this code only helps when we are used
 165          * by diff-cache --cached, which does read the cache before
 166          * calling us.
 167          */
 168         if (!active_cache)
 169                 return 0;
 170
 171         len = strlen(name);
 172         pos = cache_name_pos(name, len);
 173         if (pos < 0)
 174                 return 0;
 175         ce = active_cache[pos];
 176         if ((lstat(name, &st) < 0) ||
 177             !S_ISREG(st.st_mode) ||
 178             cache_match_stat(ce, &st) ||
 179             memcmp(sha1, ce->sha1, 20))
 180                 return 0;
 181         return 1;
 182 }
 183
 184 static void prep_temp_blob(struct diff_tempfile *temp,
 185                            void *blob,
 186                            unsigned long size,
 187                            unsigned char *sha1,
 188                            int mode)
 189 {
 190         int fd;
 191
 192         strcpy(temp->tmp_path, ".diff_XXXXXX");
 193         fd = mkstemp(temp->tmp_path);
 194         if (fd < 0)
 195                 die("unable to create temp-file");
 196         if (write(fd, blob, size) != size)
 197                 die("unable to write temp-file");
 198         close(fd);
 199         temp->name = temp->tmp_path;
 200         strcpy(temp->hex, sha1_to_hex(sha1));
 201         temp->hex[40] = 0;
 202         sprintf(temp->mode, "%06o", mode);
 203 }
 204
 205 static void prepare_temp_file(const char *name,
 206                               struct diff_tempfile *temp,
 207                               struct diff_spec *one)
 208 {
 209         static unsigned char null_sha1[20] = { 0, };
 210         int use_work_tree = 0;
 211
 212         if (!one->file_valid) {
 213         not_a_valid_file:
 214                 /* A '-' entry produces this for file-2, and
 215                  * a '+' entry produces this for file-1.
 216                  */
 217                 temp->name = "/dev/null";
 218                 strcpy(temp->hex, ".");
 219                 strcpy(temp->mode, ".");
 220                 return;
 221         }
 222
 223         if (one->sha1_valid &&
 224             (!memcmp(one->blob_sha1, null_sha1, sizeof(null_sha1)) ||
 225              work_tree_matches(name, one->blob_sha1)))
 226                 use_work_tree = 1;
 227
 228         if (!one->sha1_valid || use_work_tree) {
 229                 struct stat st;
 230                 temp->name = name;
 231                 if (lstat(temp->name, &st) < 0) {
 232                         if (errno == ENOENT)
 233                                 goto not_a_valid_file;
 234                         die("stat(%s): %s", temp->name, strerror(errno));
 235                 }
 236                 if (S_ISLNK(st.st_mode)) {
 237                         int ret;
 238                         char *buf, buf_[1024];
 239                         buf = ((sizeof(buf_) < st.st_size) ?
 240                                xmalloc(st.st_size) : buf_);
 241                         ret = readlink(name, buf, st.st_size);
 242                         if (ret < 0)
 243                                 die("readlink(%s)", name);
 244                         prep_temp_blob(temp, buf, st.st_size,
 245                                        (one->sha1_valid ?
 246                                         one->blob_sha1 : null_sha1),
 247                                        (one->sha1_valid ?
 248                                         one->mode : S_IFLNK));
 249                 }
 250                 else {
 251                         if (!one->sha1_valid)
 252                                 strcpy(temp->hex, sha1_to_hex(null_sha1));
 253                         else
 254                                 strcpy(temp->hex, sha1_to_hex(one->blob_sha1));
 255                         sprintf(temp->mode, "%06o",
 256                                 S_IFREG |ce_permissions(st.st_mode));
 257                 }
 258                 return;
 259         }
 260         else {
 261                 void *blob;
 262                 char type[20];
 263                 unsigned long size;
 264
 265                 blob = read_sha1_file(one->blob_sha1, type, &size);
 266                 if (!blob || strcmp(type, "blob"))
 267                         die("unable to read blob object for %s (%s)",
 268                             name, sha1_to_hex(one->blob_sha1));
 269                 prep_temp_blob(temp, blob, size, one->blob_sha1, one->mode);
 270                 free(blob);
 271         }
 272 }
 273
 274 static void remove_tempfile(void)
 275 {
 276         int i;
 277
 278         for (i = 0; i < 2; i++)
 279                 if (diff_temp[i].name == diff_temp[i].tmp_path) {
 280                         unlink(diff_temp[i].name);
 281                         diff_temp[i].name = NULL;
 282                 }
 283 }
 284
 285 static void remove_tempfile_on_signal(int signo)
 286 {
 287         remove_tempfile();
 288 }
 289
 290 /* An external diff command takes:
 291  *
 292  * diff-cmd name infile1 infile1-sha1 infile1-mode \
 293  *               infile2 infile2-sha1 infile2-mode.
 294  *
 295  */
 296 void run_external_diff(const char *name,
 297                        struct diff_spec *one,
 298                        struct diff_spec *two)
 299 {
 300         struct diff_tempfile *temp = diff_temp;
 301         pid_t pid;
 302         int status;
 303         static int atexit_asked = 0;
 304
 305         if (one && two) {
 306                 prepare_temp_file(name, &temp[0], one);
 307                 prepare_temp_file(name, &temp[1], two);
 308                 if (! atexit_asked &&
 309                     (temp[0].name == temp[0].tmp_path ||
 310                      temp[1].name == temp[1].tmp_path)) {
 311                         atexit_asked = 1;
 312                         atexit(remove_tempfile);
 313                 }
 314                 signal(SIGINT, remove_tempfile_on_signal);
 315         }
 316
 317         fflush(NULL);
 318         pid = fork();
 319         if (pid < 0)
 320                 die("unable to fork");
 321         if (!pid) {
 322                 const char *pgm = external_diff();
 323                 if (pgm) {
 324                         if (one && two)
 325                                 execlp(pgm, pgm,
 326                                        name,
 327                                        temp[0].name, temp[0].hex, temp[0].mode,
 328                                        temp[1].name, temp[1].hex, temp[1].mode,
 329                                        NULL);
 330                         else
 331                                 execlp(pgm, pgm, name, NULL);
 332                 }
 333                 /*
 334                  * otherwise we use the built-in one.
 335                  */
 336                 if (one && two)
 337                         builtin_diff(name, temp);
 338                 else
 339                         printf("* Unmerged path %s\n", name);
 340                 exit(0);
 341         }
 342         if (waitpid(pid, &status, 0) < 0 ||
 343             !WIFEXITED(status) || WEXITSTATUS(status)) {
 344                 /* Earlier we did not check the exit status because
 345                  * diff exits non-zero if files are different, and
 346                  * we are not interested in knowing that.  It was a
 347                  * mistake which made it harder to quit a diff-*
 348                  * session that uses the git-apply-patch-script as
 349                  * the GIT_EXTERNAL_DIFF.  A custom GIT_EXTERNAL_DIFF
 350                  * should also exit non-zero only when it wants to
 351                  * abort the entire diff-* session.
 352                  */
 353                 remove_tempfile();
 354                 fprintf(stderr, "external diff died, stopping at %s.\n", name);
 355                 exit(1);
 356         }
 357         remove_tempfile();
 358 }
 359
 360 void diff_addremove(int addremove, unsigned mode,
 361                     const unsigned char *sha1,
 362                     const char *base, const char *path)
 363 {
 364         char concatpath[PATH_MAX];
 365         struct diff_spec spec[2], *one, *two;
 366
 367         memcpy(spec[0].blob_sha1, sha1, 20);
 368         spec[0].mode = mode;
 369         spec[0].sha1_valid = spec[0].file_valid = 1;
 370         spec[1].file_valid = 0;
 371
 372         if (addremove == '+') {
 373                 one = spec + 1; two = spec;
 374         } else {
 375                 one = spec; two = one + 1;
 376         }
 377
 378         if (path) {
 379                 strcpy(concatpath, base);
 380                 strcat(concatpath, path);
 381         }
 382         run_external_diff(path ? concatpath : base, one, two);
 383 }
 384
 385 void diff_change(unsigned old_mode, unsigned new_mode,
 386                  const unsigned char *old_sha1,
 387                  const unsigned char *new_sha1,
 388                  const char *base, const char *path) {
 389         char concatpath[PATH_MAX];
 390         struct diff_spec spec[2];
 391
 392         memcpy(spec[0].blob_sha1, old_sha1, 20);
 393         spec[0].mode = old_mode;
 394         memcpy(spec[1].blob_sha1, new_sha1, 20);
 395         spec[1].mode = new_mode;
 396         spec[0].sha1_valid = spec[0].file_valid = 1;
 397         spec[1].sha1_valid = spec[1].file_valid = 1;
 398
 399         if (path) {
 400                 strcpy(concatpath, base);
 401                 strcat(concatpath, path);
 402         }
 403         run_external_diff(path ? concatpath : base, &spec[0], &spec[1]);
 404 }
 405
 406 void diff_unmerge(const char *path)
 407 {
 408         run_external_diff(path, NULL, NULL);
 409 }