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