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 <sys/mman.h>
  13 #include <string.h>
  14 #include <stdio.h>
  15 #include <ctype.h>
  16 #include <assert.h>
  17 #include "cache.h"
  18
  19 static const char git_mailsplit_usage[] =
  20 "git-mailsplit [-d<prec>] [<mbox>] <directory>";
  21
  22 static int is_from_line(const char *line, int len)
  23 {
  24         const char *colon;
  25
  26         if (len < 20 || memcmp("From ", line, 5))
  27                 return 0;
  28
  29         colon = line + len - 2;
  30         line += 5;
  31         for (;;) {
  32                 if (colon < line)
  33                         return 0;
  34                 if (*--colon == ':')
  35                         break;
  36         }
  37
  38         if (!isdigit(colon[-4]) ||
  39             !isdigit(colon[-2]) ||
  40             !isdigit(colon[-1]) ||
  41             !isdigit(colon[ 1]) ||
  42             !isdigit(colon[ 2]))
  43                 return 0;
  44
  45         /* year */
  46         if (strtol(colon+3, NULL, 10) <= 90)
  47                 return 0;
  48
  49         /* Ok, close enough */
  50         return 1;
  51 }
  52
  53 /* Could be as small as 64, enough to hold a Unix "From " line. */
  54 static char buf[4096];
  55
  56 /* Called with the first line (potentially partial)
  57  * already in buf[] -- normally that should begin with
  58  * the Unix "From " line.  Write it into the specified
  59  * file.
  60  */
  61 static int split_one(FILE *mbox, const char *name)
  62 {
  63         FILE *output = NULL;
  64         int len = strlen(buf);
  65         int fd;
  66         int status = 0;
  67
  68         if (!is_from_line(buf, len))
  69                 goto corrupt;
  70
  71         fd = open(name, O_WRONLY | O_CREAT | O_EXCL, 0666);
  72         if (fd < 0)
  73                 die("cannot open output file %s", name);
  74         output = fdopen(fd, "w");
  75
  76         /* Copy it out, while searching for a line that begins with
  77          * "From " and having something that looks like a date format.
  78          */
  79         for (;;) {
  80                 int is_partial = (buf[len-1] != '\n');
  81
  82                 if (fputs(buf, output) == EOF)
  83                         die("cannot write output");
  84
  85                 if (fgets(buf, sizeof(buf), mbox) == NULL) {
  86                         if (feof(mbox)) {
  87                                 status = 1;
  88                                 break;
  89                         }
  90                         die("cannot read mbox");
  91                 }
  92                 len = strlen(buf);
  93                 if (!is_partial && is_from_line(buf, len))
  94                         break; /* done with one message */
  95         }
  96         fclose(output);
  97         return status;
  98
  99  corrupt:
 100         if (output)
 101                 fclose(output);
 102         unlink(name);
 103         fprintf(stderr, "corrupt mailbox\n");
 104         exit(1);
 105 }
 106
 107 int main(int argc, const char **argv)
 108 {
 109         int i, nr, nr_prec = 4;
 110         FILE *mbox = NULL;
 111
 112         for (i = 1; i < argc; i++) {
 113                 const char *arg = argv[i];
 114
 115                 if (arg[0] != '-')
 116                         break;
 117                 /* do flags here */
 118                 if (!strncmp(arg, "-d", 2)) {
 119                         nr_prec = strtol(arg + 2, NULL, 10);
 120                         if (nr_prec < 3 || 10 <= nr_prec)
 121                                 usage(git_mailsplit_usage);
 122                         continue;
 123                 }
 124         }
 125
 126         /* Either one remaining arg (dir), or two (mbox and dir) */
 127         switch (argc - i) {
 128         case 1:
 129                 mbox = stdin;
 130                 break;
 131         case 2:
 132                 if ((mbox = fopen(argv[i], "r")) == NULL)
 133                         die("cannot open mbox %s for reading", argv[i]);
 134                 break;
 135         default:
 136                 usage(git_mailsplit_usage);
 137         }
 138         if (chdir(argv[argc - 1]) < 0)
 139                 usage(git_mailsplit_usage);
 140
 141         nr = 0;
 142         if (fgets(buf, sizeof(buf), mbox) == NULL)
 143                 die("cannot read mbox");
 144
 145         for (;;) {
 146                 char name[10];
 147
 148                 sprintf(name, "%0*d", nr_prec, ++nr);
 149                 switch (split_one(mbox, name)) {
 150                 case 0:
 151                         break;
 152                 case 1:
 153                         printf("%d\n", nr);
 154                         return 0;
 155                 default:
 156                         exit(1);
 157                 }
 158         }
 159 }