h2chin

   1 #!/usr/bin/env perl
   2
   3 # h2chin
   4 # File ID: c9650f12-5d3a-11df-aa5c-90e6ba3022ac
   5 # Konverterer fra numeriske entities til oversatt kinesisk.
   6
   7 use strict;
   8 use warnings;
   9
  10 my %Uh;
  11
  12 $| = 1;
  13
  14 open(FromFP, "</tmp/Unihan-def.txt") || die("Klarte ikke å åpne unihan for lesing: $!");
  15 while(<FromFP>) {
  16         if (/^U\+(....)\s+kDefinition\s+(.*)/) {
  17                 $Uh{$1} = $2;
  18                 # print (STDERR "$1\n");
  19         } else {
  20                 # print(STDERR "Bomma på $_\n");
  21         }
  22 }
  23
  24 close(FromFP);
  25
  26 while (<>) {
  27         s!&#x([0-9A-F][0-9A-F][0-9A-F][0-9A-F]);!sprintf("\n[%s]", length($Uh{$1}) ? $Uh{$1} : join("", "U+", $1))!gei;
  28         print;
  29 }