builtin-mailsplit.c

   1 /*
   2  * Totally braindamaged mbox splitter program.
   3  *
   4  * It just splits a mbox into a list of files: "0001" "0002" ..
   5  * so you can process them further from there.
   6  */
   7 #include "cache.h"
   8 #include "builtin.h"
   9 #include "path-list.h"
  10
  11 static const char git_mailsplit_usage[] =
  12 "git-mailsplit [-d<prec>] [-f<n>] [-b] -o<directory> <mbox>|<Maildir>...";
  13
  14 static int is_from_line(const char *line, int len)
  15 {
  16         const char *colon;
  17
  18         if (len < 20 || memcmp("From ", line, 5))
  19                 return 0;
  20
  21         colon = line + len - 2;
  22         line += 5;
  23         for (;;) {
  24                 if (colon < line)
  25                         return 0;
  26                 if (*--colon == ':')
  27                         break;
  28         }
  29
  30         if (!isdigit(colon[-4]) ||
  31             !isdigit(colon[-2]) ||
  32             !isdigit(colon[-1]) ||
  33             !isdigit(colon[ 1]) ||
  34             !isdigit(colon[ 2]))
  35                 return 0;
  36
  37         /* year */
  38         if (strtol(colon+3, NULL, 10) <= 90)
  39                 return 0;
  40
  41         /* Ok, close enough */
  42         return 1;
  43 }
  44
  45 /* Could be as small as 64, enough to hold a Unix "From " line. */
  46 static char buf[4096];
  47
  48 /* We cannot use fgets() because our lines can contain NULs */
  49 int read_line_with_nul(char *buf, int size, FILE *in)
  50 {
  51         int len = 0, c;
  52
  53         for (;;) {
  54                 c = getc(in);
  55                 buf[len++] = c;
  56                 if (c == EOF || c == '\n' || len + 1 >= size)
  57                         break;
  58         }
  59
  60         if (c == EOF)
  61                 len--;
  62         buf[len] = '\0';
  63
  64         return len;
  65 }
  66
  67 /* Called with the first line (potentially partial)
  68  * already in buf[] -- normally that should begin with
  69  * the Unix "From " line.  Write it into the specified
  70  * file.
  71  */
  72 static int split_one(FILE *mbox, const char *name, int allow_bare)
  73 {
  74         FILE *output = NULL;
  75         int len = strlen(buf);
  76         int fd;
  77         int status = 0;
  78         int is_bare = !is_from_line(buf, len);
  79
  80         if (is_bare && !allow_bare)
  81                 goto corrupt;
  82
  83         fd = open(name, O_WRONLY | O_CREAT | O_EXCL, 0666);
  84         if (fd < 0)
  85                 die("cannot open output file %s", name);
  86         output = fdopen(fd, "w");
  87
  88         /* Copy it out, while searching for a line that begins with
  89          * "From " and having something that looks like a date format.
  90          */
  91         for (;;) {
  92                 int is_partial = len && buf[len-1] != '\n';
  93
  94                 if (fwrite(buf, 1, len, output) != len)
  95                         die("cannot write output");
  96
  97                 len = read_line_with_nul(buf, sizeof(buf), mbox);
  98                 if (len == 0) {
  99                         if (feof(mbox)) {
 100                                 status = 1;
 101                                 break;
 102                         }
 103                         die("cannot read mbox");
 104                 }
 105                 if (!is_partial && !is_bare && is_from_line(buf, len))
 106                         break; /* done with one message */
 107         }
 108         fclose(output);
 109         return status;
 110
 111  corrupt:
 112         if (output)
 113                 fclose(output);
 114         unlink(name);
 115         fprintf(stderr, "corrupt mailbox\n");
 116         exit(1);
 117 }
 118
 119 static int populate_maildir_list(struct path_list *list, const char *path)
 120 {
 121         DIR *dir;
 122         struct dirent *dent;
 123         char name[PATH_MAX];
 124         char *subs[] = { "cur", "new", NULL };
 125         char **sub;
 126
 127         for (sub = subs; *sub; ++sub) {
 128                 snprintf(name, sizeof(name), "%s/%s", path, *sub);
 129                 if ((dir = opendir(name)) == NULL) {
 130                         if (errno == ENOENT)
 131                                 continue;
 132                         error("cannot opendir %s (%s)", name, strerror(errno));
 133                         return -1;
 134                 }
 135
 136                 while ((dent = readdir(dir)) != NULL) {
 137                         if (dent->d_name[0] == '.')
 138                                 continue;
 139                         snprintf(name, sizeof(name), "%s/%s", *sub, dent->d_name);
 140                         path_list_insert(name, list);
 141                 }
 142
 143                 closedir(dir);
 144         }
 145
 146         return 0;
 147 }
 148
 149 static int split_maildir(const char *maildir, const char *dir,
 150         int nr_prec, int skip)
 151 {
 152         char file[PATH_MAX];
 153         char name[PATH_MAX];
 154         int ret = -1;
 155         int i;
 156         struct path_list list = {NULL, 0, 0, 1};
 157
 158         if (populate_maildir_list(&list, maildir) < 0)
 159                 goto out;
 160
 161         for (i = 0; i < list.nr; i++) {
 162                 FILE *f;
 163                 snprintf(file, sizeof(file), "%s/%s", maildir, list.items[i].path);
 164                 f = fopen(file, "r");
 165                 if (!f) {
 166                         error("cannot open mail %s (%s)", file, strerror(errno));
 167                         goto out;
 168                 }
 169
 170                 if (fgets(buf, sizeof(buf), f) == NULL) {
 171                         error("cannot read mail %s (%s)", file, strerror(errno));
 172                         goto out;
 173                 }
 174
 175                 sprintf(name, "%s/%0*d", dir, nr_prec, ++skip);
 176                 split_one(f, name, 1);
 177
 178                 fclose(f);
 179         }
 180
 181         ret = skip;
 182 out:
 183         path_list_clear(&list, 1);
 184         return ret;
 185 }
 186
 187 static int split_mbox(const char *file, const char *dir, int allow_bare,
 188                       int nr_prec, int skip)
 189 {
 190         char name[PATH_MAX];
 191         int ret = -1;
 192         int peek;
 193
 194         FILE *f = !strcmp(file, "-") ? stdin : fopen(file, "r");
 195         int file_done = 0;
 196
 197         if (!f) {
 198                 error("cannot open mbox %s", file);
 199                 goto out;
 200         }
 201
 202         do {
 203                 peek = fgetc(f);
 204         } while (isspace(peek));
 205         ungetc(peek, f);
 206
 207         if (fgets(buf, sizeof(buf), f) == NULL) {
 208                 /* empty stdin is OK */
 209                 if (f != stdin) {
 210                         error("cannot read mbox %s", file);
 211                         goto out;
 212                 }
 213                 file_done = 1;
 214         }
 215
 216         while (!file_done) {
 217                 sprintf(name, "%s/%0*d", dir, nr_prec, ++skip);
 218                 file_done = split_one(f, name, allow_bare);
 219         }
 220
 221         if (f != stdin)
 222                 fclose(f);
 223
 224         ret = skip;
 225 out:
 226         return ret;
 227 }
 228
 229 int cmd_mailsplit(int argc, const char **argv, const char *prefix)
 230 {
 231         int nr = 0, nr_prec = 4, num = 0;
 232         int allow_bare = 0;
 233         const char *dir = NULL;
 234         const char **argp;
 235         static const char *stdin_only[] = { "-", NULL };
 236
 237         for (argp = argv+1; *argp; argp++) {
 238                 const char *arg = *argp;
 239
 240                 if (arg[0] != '-')
 241                         break;
 242                 /* do flags here */
 243                 if ( arg[1] == 'd' ) {
 244                         nr_prec = strtol(arg+2, NULL, 10);
 245                         if (nr_prec < 3 || 10 <= nr_prec)
 246                                 usage(git_mailsplit_usage);
 247                         continue;
 248                 } else if ( arg[1] == 'f' ) {
 249                         nr = strtol(arg+2, NULL, 10);
 250                 } else if ( arg[1] == 'b' && !arg[2] ) {
 251                         allow_bare = 1;
 252                 } else if ( arg[1] == 'o' && arg[2] ) {
 253                         dir = arg+2;
 254                 } else if ( arg[1] == '-' && !arg[2] ) {
 255                         argp++; /* -- marks end of options */
 256                         break;
 257                 } else {
 258                         die("unknown option: %s", arg);
 259                 }
 260         }
 261
 262         if ( !dir ) {
 263                 /* Backwards compatibility: if no -o specified, accept
 264                    <mbox> <dir> or just <dir> */
 265                 switch (argc - (argp-argv)) {
 266                 case 1:
 267                         dir = argp[0];
 268                         argp = stdin_only;
 269                         break;
 270                 case 2:
 271                         stdin_only[0] = argp[0];
 272                         dir = argp[1];
 273                         argp = stdin_only;
 274                         break;
 275                 default:
 276                         usage(git_mailsplit_usage);
 277                 }
 278         } else {
 279                 /* New usage: if no more argument, parse stdin */
 280                 if ( !*argp )
 281                         argp = stdin_only;
 282         }
 283
 284         while (*argp) {
 285                 const char *arg = *argp++;
 286                 struct stat argstat;
 287                 int ret = 0;
 288
 289                 if (arg[0] == '-' && arg[1] == 0) {
 290                         ret = split_mbox(arg, dir, allow_bare, nr_prec, nr);
 291                         if (ret < 0) {
 292                                 error("cannot split patches from stdin");
 293                                 return 1;
 294                         }
 295                         num += (ret - nr);
 296                         nr = ret;
 297                         continue;
 298                 }
 299
 300                 if (stat(arg, &argstat) == -1) {
 301                         error("cannot stat %s (%s)", arg, strerror(errno));
 302                         return 1;
 303                 }
 304
 305                 if (S_ISDIR(argstat.st_mode))
 306                         ret = split_maildir(arg, dir, nr_prec, nr);
 307                 else
 308                         ret = split_mbox(arg, dir, allow_bare, nr_prec, nr);
 309
 310                 if (ret < 0) {
 311                         error("cannot split patches from %s", arg);
 312                         return 1;
 313                 }
 314                 num += (ret - nr);
 315                 nr = ret;
 316         }
 317
 318         printf("%d\n", num);
 319
 320         return 0;
 321 }