Convert a hex-string into binary file with plain awk under cygwin

Question

Trying to convert a hexstring given as string in the form 31ff097112 with plain awk into a binary file using cygwin (Works when run in linux). Tried the following:

BEGIN {
  BINMODE=3;
  ORS="";
  s="000102030405060708ff800f0e0d0c0b0a";
  # not working on cygwin for values >=0x80:
  len=length(s);
  print "">"broken.bin"
  for(i=1; i<=len; i+=2) {
    printf("%c", strtonum("0x"substr(s, i, 2)))>>"broken.bin";
  }
  #working, but uses external xxd command:
  print s>"/tmp/xxx";
  system ("cat /tmp/xxx|xxd -r -p>good.bin");
  exit;
}

So the problem with my plain awk solution is, that all values >= 0x80 are somehow written utf8-encoded to the resulting file:

$ hexdump -C broken.bin
00000000  00 01 02 03 04 05 06 07  08 c3 bf c2 80 0f 0e 0d  |................|
00000010  0c 0b 0a                                          |...|
00000013

whereas the xxd-tool (Found a hint there) does it right (as expected :-):

00000000  00 01 02 03 04 05 06 07  08 ff 80 0f 0e 0d 0c 0b  |................|
00000010  0a                                                |.|

So I would like to know wether there is a solution which works with awk (no, not perl, not python, not C) only using awk's features.

NB: My plain awk solutions works with linux awk perfect, so maybe there are some other options to be set within cygwin to make that work?

Answer 1

Sorry to answer my own question, just found the solution: invoke awk with -b or --characters-as-bytes, then it also works from within cygwin...