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