[patch] Adding Numerical Suffixes to Split

[Capt Jesse Kornblum USAF]

Sorry it's been so long since I've written; the Air Force moved me in 
the past month.

I'm still interested in adding support for numerical suffixes to split.

Several people have commented that the change from alphabetic suffixes 
to numerical ones can be accomplished via a shell script. That's true, 
but my question to you, is: Why? Why should the user have to remember a 
long line of arcane shell commands instead of just adding a single 
command line flag that already exists in a sister program (csplit)?

The numerical suffix is *the* standard for forensic images (think dd 
files, not graphics). Originally developed for a forensic imaging 
program called Safeback in late 80's, every forensic examination program 
now uses numerical suffixes to identity parts of a single image. Such 
programs include EnCase (http://encase.com/), iLook 
(http://www.ilook-forensics.org/), and Autopsy (http://www.sleuthkit.org/).

Your concern for preserving the split standard is admirable, but I don't 
think that this patch will break that standard. That is, files created 
with numerical suffixes can be read in the same manner as files created 
with alphabetic suffixes. Here's an example with a small text file 
called "bar." I split the 'bar' file using regular split and then my 
modified version. Both can then be 'cat'ed back together to be the same 
file.

$ ls -l bar
-rw-r--r--    1 jessek   users        5120 Aug  7 15:34 bar
$ split -b 500 bar normal
$ /home/jessek/coreutils-5.0/src/split -n -b 500 bar digits
$ ls normal* digits*
digits01  digits04  digits07  digits10  normalab  normalae  normalah 
normalak
digits02  digits05  digits08  digits11  normalac  normalaf  normalai
digits03  digits06  digits09  normalaa  normalad  normalag  normalaj
$ cat normal* > all-normal
$ cat digits* > all-digits
$ diff all-normal all-digits
$

Thus, any set of files created with numerical suffixes is still 
compatible with any set of files created with alphabetic suffixes. I'm 
open to argument on this point, of course. :)

Somebody asked if csplit would work for our purposes. Unfortunately we 
need a program that splits files based on their size. csplit looks only 
at a file's content.

Below, as requested, is the formatted patch for split to allow numerical 
suffixes. (I had some CVS issues, but think I got it right.)

2003-08-08  Jesse Kornblum  <address@hidden>

       Add support for numerical suffixes in split

       * src/split.c - Add support for -n

Index: patch.c
===================================================================
RCS file: /cvsroot/coreutils/coreutils/patch,v
retrieving revision x.x
diff -p -u -rx.x patch.c
--- split-orig.c        Thu Aug  7 15:00:25 2003
+++ split.c     Thu Aug  7 15:11:11 2003
@@ -63,6 +63,9 @@ static size_t suffix_length = DEFAULT_SU
/* Name of input file.  May be "-".  */
static char *infile;

+/* If non-zero, use numeric suffixes instead of characters */
+static int suffix_type;
+
/* Descriptor on which input file is open.  */
static int input_desc;

@@ -78,6 +81,7 @@ static struct option const longopts[] =
  {"bytes", required_argument, NULL, 'b'},
  {"lines", required_argument, NULL, 'l'},
  {"line-bytes", required_argument, NULL, 'C'},
+  {"numbers", no_argument, &suffix_type, 'n'},
  {"suffix-length", required_argument, NULL, 'a'},
  {"verbose", no_argument, &verbose, 0},
  {GETOPT_HELP_OPTION_DECL},
@@ -109,6 +113,7 @@ Mandatory arguments to long options are
  -a, --suffix-length=N   use suffixes of length N (default %d)\n\
  -b, --bytes=SIZE        put SIZE bytes per output file\n\
  -C, --line-bytes=SIZE   put at most SIZE bytes of lines per output 
file\n\
+  -n, --numbers           use digits for suffixes instead of numbers\n\
  -l, --lines=NUMBER      put NUMBER lines per output file\n\
"), DEFAULT_SUFFIX_LENGTH);
      fputs (_("\
@@ -143,7 +148,12 @@ next_file_name (void)
      outfile = xmalloc (outfile_length + 1);
      outfile_mid = outfile + outbase_length;
      memcpy (outfile, outbase, outbase_length);
-      memset (outfile_mid, 'a', suffix_length);
+      if (!suffix_type)
+       memset (outfile_mid, 'a', suffix_length);
+      else {
+       memset (outfile_mid, '0', suffix_length);
+       outfile_mid[suffix_length - 1] = '1';
+      }
      outfile[outfile_length] = 0;

#if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
@@ -165,10 +175,17 @@ next_file_name (void)
      /* Increment the suffix in place, if possible.  */
      char *p;
-      for (p = outfile_mid + suffix_length; outfile_mid < p; *--p = 'a')
-       if (p[-1]++ != 'z')
-         return;
-      error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
+      if (!suffix_type) {
+       for (p = outfile_mid + suffix_length; outfile_mid < p; *--p = 'a')
+         if (p[-1]++ != 'z')
+           return;
+       error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
+      } else {
+       for (p = outfile_mid + suffix_length; outfile_mid < p; *--p = '0')
+         if (p[-1]++ != '9')
+           return;
+       error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
+      }
    }
}

@@ -376,7 +393,7 @@ main (int argc, char **argv)
      int this_optind = optind ? optind : 1;
      long int tmp_long;

-      c = getopt_long (argc, argv, "0123456789C:a:b:l:", longopts, NULL);
+      c = getopt_long (argc, argv, "0123456789nC:a:b:l:", longopts, NULL);
      if (c == -1)
       break;

@@ -385,6 +402,10 @@ main (int argc, char **argv)
       case 0:
         break;

+       case 'n':
+         suffix_type = 1;
+         break;
+
       case 'a':
         {
           unsigned long tmp;

--
Jesse Kornblum, Capt, USAF
United States Naval Academy
Chauvenet Room 329
572 Holloway Rd. Stop 9F
Annapolis, MD 21402-5002
Comm 410-293-6821  DSN 281-6821
Fax 410-293-2686  Fax DSN 281-2686
e-mail: address@hidden
http://www.cs.usna.edu/~kornblum/