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 <unistd.h>
   8 #include <stdlib.h>
   9 #include <fcntl.h>
  10 #include <sys/types.h>
  11 #include <sys/stat.h>
  12 #include <string.h>
  13 #include <stdio.h>
  14 #include <ctype.h>
  15 #include <assert.h>
  16 #include "cache.h"
  17
  18 static const char git_mailsplit_usage[] =
  19 "git-mailsplit [-d<prec>] [<mbox>] <directory>";
  20
  21 static int is_from_line(const char *line, int len)
  22 {
  23         const char *colon;
  24
  25         if (len < 20 || memcmp("From ", line, 5))
  26                 return 0;
  27
  28         colon = line + len - 2;
  29         line += 5;
  30         for (;;) {
  31                 if (colon < line)
  32                         return 0;
  33                 if (*--colon == ':')
  34                         break;
  35         }
  36
  37         if (!isdigit(colon[-4]) ||
  38             !isdigit(colon[-2]) ||
  39             !isdigit(colon[-1]) ||
  40             !isdigit(colon[ 1]) ||
  41             !isdigit(colon[ 2]))
  42                 return 0;
  43
  44         /* year */
  45         if (strtol(colon+3, NULL, 10) <= 90)
  46                 return 0;
  47
  48         /* Ok, close enough */
  49         return 1;
  50 }
  51
  52 /* Could be as small as 64, enough to hold a Unix "From " line. */
  53 static char buf[4096];
  54
  55 /* Called with the first line (potentially partial)
  56  * already in buf[] -- normally that should begin with
  57  * the Unix "From " line.  Write it into the specified
  58  * file.
  59  */
  60 static int split_one(FILE *mbox, const char *name)
  61 {
  62         FILE *output = NULL;
  63         int len = strlen(buf);
  64         int fd;
  65         int status = 0;
  66
  67         if (!is_from_line(buf, len))
  68                 goto corrupt;
  69
  70         fd = open(name, O_WRONLY | O_CREAT | O_EXCL, 0666);
  71         if (fd < 0)
  72                 die("cannot open output file %s", name);
  73         output = fdopen(fd, "w");
  74
  75         /* Copy it out, while searching for a line that begins with
  76          * "From " and having something that looks like a date format.
  77          */
  78         for (;;) {
  79                 int is_partial = (buf[len-1] != '\n');
  80
  81                 if (fputs(buf, output) == EOF)
  82                         die("cannot write output");
  83
  84                 if (fgets(buf, sizeof(buf), mbox) == NULL) {
  85                         if (feof(mbox)) {
  86                                 status = 1;
  87                                 break;
  88                         }
  89                         die("cannot read mbox");
  90                 }
  91                 len = strlen(buf);
  92                 if (!is_partial && is_from_line(buf, len))
  93                         break; /* done with one message */
  94         }
  95         fclose(output);
  96         return status;
  97
  98  corrupt:
  99         if (output)
 100                 fclose(output);
 101         unlink(name);
 102         fprintf(stderr, "corrupt mailbox\n");
 103         exit(1);
 104 }
 105
 106 int main(int argc, const char **argv)
 107 {
 108         int i, nr, nr_prec = 4;
 109         FILE *mbox = NULL;
 110
 111         for (i = 1; i < argc; i++) {
 112                 const char *arg = argv[i];
 113
 114                 if (arg[0] != '-')
 115                         break;
 116                 /* do flags here */
 117                 if (!strncmp(arg, "-d", 2)) {
 118                         nr_prec = strtol(arg + 2, NULL, 10);
 119                         if (nr_prec < 3 || 10 <= nr_prec)
 120                                 usage(git_mailsplit_usage);
 121                         continue;
 122                 }
 123         }
 124
 125         /* Either one remaining arg (dir), or two (mbox and dir) */
 126         switch (argc - i) {
 127         case 1:
 128                 mbox = stdin;
 129                 break;
 130         case 2:
 131                 if ((mbox = fopen(argv[i], "r")) == NULL)
 132                         die("cannot open mbox %s for reading", argv[i]);
 133                 break;
 134         default:
 135                 usage(git_mailsplit_usage);
 136         }
 137         if (chdir(argv[argc - 1]) < 0)
 138                 usage(git_mailsplit_usage);
 139
 140         nr = 0;
 141         if (fgets(buf, sizeof(buf), mbox) == NULL)
 142                 die("cannot read mbox");
 143
 144         for (;;) {
 145                 char name[10];
 146
 147                 sprintf(name, "%0*d", nr_prec, ++nr);
 148                 switch (split_one(mbox, name)) {
 149                 case 0:
 150                         break;
 151                 case 1:
 152                         printf("%d\n", nr);
 153                         return 0;
 154                 default:
 155                         exit(1);
 156                 }
 157         }
 158 }