Update CJK converters and their generating scripts

scripts/big5_gen.sh

 #!/bin/sh
 # Copyright 2015 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 # References:
 #   https://encoding.spec.whatwg.org/#big5
 
 # This script downloads the following file.
 #   https://encoding.spec.whatwg.org/index-big5.txt
@@ skipping 11 matching lines @@
 # ***************************************************************************
 <code_set_name>               "big5-html"
 <char_name_mask>              "AXXXX"
 <mb_cur_max>                  2
 <mb_cur_min>                  1
 <uconv_class>                 "MBCS"
 <subchar>                     \x3F
 <icu:charsetFamily>           "ASCII"
 
 # 'p' is for the range that may produce non-BMP code points.
+# 'i' is to make the code range illegal.
+# Big5 has a lot of small holes in the 2nd byte. If it's in the ASCII range,
+# the 2nd byte has to be added back to the stream to be compliant to the
+# encoding spec. Each state adds 1kB in the data size.
 # See http://userguide.icu-project.org/conversion/data.
-<icu:state>                   0-7f, 87-fe:1, 87-a0:2, c8:2, fa-fe:2
+<icu:state>                   0-7f, a1-fe:1, 87-a0:2, c8:2, fa-fe:2, 87:3, 89:4, 8a:5, 8b:6, 8d:7, 9b:8, 9f:9, a0:a
 <icu:state>                   40-7e, a1-fe
 <icu:state>                   40-7e.p, a1-fe.p
+<icu:state>                   40-7e.p, a1-fe.p, 66.i
+<icu:state>                   40-7e.p, a1-fe.p, 42.i, 44.i, 45.i, 4a-4b.i
+<icu:state>                   40-7e.p, a1-fe.p, 42.i, 63.i, 75.i
+<icu:state>                   40-7e.p, a1-fe.p, 54.i
+<icu:state>                   40-7e.p, a1-fe.p, 41.i
+<icu:state>                   40-7e.p, a1-fe.p, 61.i
+<icu:state>                   40-7e.p, a1-fe.p, 4e.i
+<icu:state>                   40-7e.p, a1-fe.p, 54.i, 57.i, 5a.i, 62.i, 72.i
 
 CHARMAP
 PREAMBLE
 }
 
 function ascii {
   for i in $(seq 0 127)
   do
     printf '<U%04X> \\x%02X |0\n' $i $i
   done
@@ skipping 34 matching lines @@
 function unsorted_table {
   two_char_seq
   big5
 }
 
 wget -N -r -nd https://encoding.spec.whatwg.org/index-big5.txt
 preamble
 ascii
 unsorted_table | sort -k4  | uniq | cut -f 1-3 -d ' '
 echo 'END CHARMAP'