One Level Up Top Level

contrib/unicode2nginx/unicode-to-nginx.pl - nginx source code

  1. #!/usr/bin/perl -w

  3. # Convert unicode mappings to nginx configuration file format.

  5. # You may find useful mappings in various places, including
  6. # unicode.org official site:
  7. #
  8. # http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP1251.TXT
  9. # http://www.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8-R.TXT

  11. # Needs perl 5.6 or later.

  13. # Written by Maxim Dounin, mdounin@mdounin.ru

  15. ###############################################################################

  17. require 5.006;

  19. while (<>) {
  20.     # Skip comments and empty lines

  22.     next if /^#/;
  23.     next if /^\s*$/;
  24.     chomp;

  26.     # Convert mappings

  28.     if (/^\s*0x(..)\s*0x(....)\s*(#.*)/) {
  29.         # Mapping <from-code> <unicode-code> "#" <unicode-name>
  30.         my $cs_code = $1;
  31.         my $un_code = $2;
  32.         my $un_name = $3;

  34.         # Produce UTF-8 sequence from character code;

  36.         my $un_utf8 = join('',
  37.             map { sprintf("%02X", $_) }
  38.             unpack("U0C*", pack("U", hex($un_code)))
  39.         );

  41.         print "    $cs_code  $un_utf8 ; $un_name\n";

  43.     } else {
  44.         warn "Unrecognized line: '$_'";
  45.     }
  46. }

  48. ###############################################################################

One Level Up Top Level