Make all the single byte encodings compliant to the encoding spec.

source/data/mappings/noop-gb2312_gl.ucm

Index: source/data/mappings/noop-gb2312_gl.ucm
===================================================================
--- source/data/mappings/noop-gb2312_gl.ucm	(revision 291619)
+++ source/data/mappings/noop-gb2312_gl.ucm	(working copy)
@@ -1,32 +0,0 @@
-#________________________________________________________________________
-#
-# (C) COPYRIGHT International Business Machines Corp. 2009
-#     All Rights Reserved
-#
-#________________________________________________________________________
-#
-#       Name:    "Fake" Unicode to ISO-IR-165 table
-#       Date:    December 2, 2009
-#
-# This is not a real GB2312 GL table, but a "fake" table to return U+FFFD
-# for every byte sequence valid in GB2312 GL. Chrome and Android
-# do not want to include the real table for ISO-1022-CN(-Ext) that are extremely
-# rare, if used at all. However, the lack of support gets
-# them vulernable to a security issue. By including this 'fake' table
-# and two others, they can be protected while increasing the ICU data size
-# by ~ 10kB.
-<code_set_name>               "noop-gb2312_gl"
-<char_name_mask>              "AXXXX"
-<mb_cur_max>                  2
-<mb_cur_min>                  2
-<uconv_class>                 "DBCS"
-<subchar>                     \x21\x21
-<icu:charsetFamily>           "ASCII"
-<icu:alias>                   "ibm-5478_STD"
-
-<icu:state>                   0-20:2, 21-7e:1, 7f-ff:2
-<icu:state>                   21-7e.u
-<icu:state>
-
-CHARMAP
-END CHARMAP