I need to write a float in big-endian format on a little-endian machine. The only way I found to do it is to pack the number into float format, rip apart the bytes, reverse them, and repack them.

This works fine, but I'm little dizzy from hoop-jumping. Is there a shorter way to get there? Here is my example, in baby Perl to be clear about the steps. Please note: I'm not looking for denser code, I know I can use map to eliminate temporary variables. I'm looking for a better algorithm.

PackFoo

#!/usr/bin/perl

  use strict;

  my ($num, $str, @bytes);

  $num = -1.18024688796416e+29;

  $str = pack('f', $num);

  showstring('native', $str);

  @bytes = split('', $str);
  @bytes = reverse(@bytes);

  $str = join('', @bytes);
  $str = pack('a4', $str);

  showstring('alien ', $str);

#-----------------------------------------------------------
sub showstring {

  my ($label, $str) = @_;

  my @bytes = unpack('H2H2H2H2', $str);

  print "$label 0x", join("", @bytes), "\n";
}
[download]