You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
177 lines
8.1 KiB
JavaScript
177 lines
8.1 KiB
JavaScript
10 months ago
|
"use strict";
|
||
|
|
|
||
|
|
// Description of supported double byte encodings and aliases.
|
||
|
|
// Tables are not require()-d until they are needed to speed up library load.
|
||
|
|
// require()-s are direct to support Browserify.
|
||
|
|
|
||
|
|
module.exports = {
|
||
|
|
|
||
|
|
// == Japanese/ShiftJIS ====================================================
|
||
|
|
// All japanese encodings are based on JIS X set of standards:
|
||
|
|
// JIS X 0201 - Single-byte encoding of ASCII + ¥ + Kana chars at 0xA1-0xDF.
|
||
|
|
// JIS X 0208 - Main set of 6879 characters, placed in 94x94 plane, to be encoded by 2 bytes.
|
||
|
|
// Has several variations in 1978, 1983, 1990 and 1997.
|
||
|
|
// JIS X 0212 - Supplementary plane of 6067 chars in 94x94 plane. 1990. Effectively dead.
|
||
|
|
// JIS X 0213 - Extension and modern replacement of 0208 and 0212. Total chars: 11233.
|
||
|
|
// 2 planes, first is superset of 0208, second - revised 0212.
|
||
|
|
// Introduced in 2000, revised 2004. Some characters are in Unicode Plane 2 (0x2xxxx)
|
||
|
|
|
||
|
|
// Byte encodings are:
|
||
|
|
// * Shift_JIS: Compatible with 0201, uses not defined chars in top half as lead bytes for double-byte
|
||
|
|
// encoding of 0208. Lead byte ranges: 0x81-0x9F, 0xE0-0xEF; Trail byte ranges: 0x40-0x7E, 0x80-0x9E, 0x9F-0xFC.
|
||
|
|
// Windows CP932 is a superset of Shift_JIS. Some companies added more chars, notably KDDI.
|
||
|
|
// * EUC-JP: Up to 3 bytes per character. Used mostly on *nixes.
|
||
|
|
// 0x00-0x7F - lower part of 0201
|
||
|
|
// 0x8E, 0xA1-0xDF - upper part of 0201
|
||
|
|
// (0xA1-0xFE)x2 - 0208 plane (94x94).
|
||
|
|
// 0x8F, (0xA1-0xFE)x2 - 0212 plane (94x94).
|
||
|
|
// * JIS X 208: 7-bit, direct encoding of 0208. Byte ranges: 0x21-0x7E (94 values). Uncommon.
|
||
|
|
// Used as-is in ISO2022 family.
|
||
|
|
// * ISO2022-JP: Stateful encoding, with escape sequences to switch between ASCII,
|
||
|
|
// 0201-1976 Roman, 0208-1978, 0208-1983.
|
||
|
|
// * ISO2022-JP-1: Adds esc seq for 0212-1990.
|
||
|
|
// * ISO2022-JP-2: Adds esc seq for GB2313-1980, KSX1001-1992, ISO8859-1, ISO8859-7.
|
||
|
|
// * ISO2022-JP-3: Adds esc seq for 0201-1976 Kana set, 0213-2000 Planes 1, 2.
|
||
|
|
// * ISO2022-JP-2004: Adds 0213-2004 Plane 1.
|
||
|
|
//
|
||
|
|
// After JIS X 0213 appeared, Shift_JIS-2004, EUC-JISX0213 and ISO2022-JP-2004 followed, with just changing the planes.
|
||
|
|
//
|
||
|
|
// Overall, it seems that it's a mess :( http://www8.plala.or.jp/tkubota1/unicode-symbols-map2.html
|
||
|
|
|
||
|
|
'shiftjis': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/shiftjis.json') },
|
||
|
|
encodeAdd: {'\u00a5': 0x5C, '\u203E': 0x7E},
|
||
|
|
encodeSkipVals: [{from: 0xED40, to: 0xF940}],
|
||
|
|
},
|
||
|
|
'csshiftjis': 'shiftjis',
|
||
|
|
'mskanji': 'shiftjis',
|
||
|
|
'sjis': 'shiftjis',
|
||
|
|
'windows31j': 'shiftjis',
|
||
|
|
'ms31j': 'shiftjis',
|
||
|
|
'xsjis': 'shiftjis',
|
||
|
|
'windows932': 'shiftjis',
|
||
|
|
'ms932': 'shiftjis',
|
||
|
|
'932': 'shiftjis',
|
||
|
|
'cp932': 'shiftjis',
|
||
|
|
|
||
|
|
'eucjp': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/eucjp.json') },
|
||
|
|
encodeAdd: {'\u00a5': 0x5C, '\u203E': 0x7E},
|
||
|
|
},
|
||
|
|
|
||
|
|
// TODO: KDDI extension to Shift_JIS
|
||
|
|
// TODO: IBM CCSID 942 = CP932, but F0-F9 custom chars and other char changes.
|
||
|
|
// TODO: IBM CCSID 943 = Shift_JIS = CP932 with original Shift_JIS lower 128 chars.
|
||
|
|
|
||
|
|
|
||
|
|
// == Chinese/GBK ==========================================================
|
||
|
|
// http://en.wikipedia.org/wiki/GBK
|
||
|
|
// We mostly implement W3C recommendation: https://www.w3.org/TR/encoding/#gbk-encoder
|
||
|
|
|
||
|
|
// Oldest GB2312 (1981, ~7600 chars) is a subset of CP936
|
||
|
|
'gb2312': 'cp936',
|
||
|
|
'gb231280': 'cp936',
|
||
|
|
'gb23121980': 'cp936',
|
||
|
|
'csgb2312': 'cp936',
|
||
|
|
'csiso58gb231280': 'cp936',
|
||
|
|
'euccn': 'cp936',
|
||
|
|
|
||
|
|
// Microsoft's CP936 is a subset and approximation of GBK.
|
||
|
|
'windows936': 'cp936',
|
||
|
|
'ms936': 'cp936',
|
||
|
|
'936': 'cp936',
|
||
|
|
'cp936': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp936.json') },
|
||
|
|
},
|
||
|
|
|
||
|
|
// GBK (~22000 chars) is an extension of CP936 that added user-mapped chars and some other.
|
||
|
|
'gbk': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp936.json').concat(require('./tables/gbk-added.json')) },
|
||
|
|
},
|
||
|
|
'xgbk': 'gbk',
|
||
|
|
'isoir58': 'gbk',
|
||
|
|
|
||
|
|
// GB18030 is an algorithmic extension of GBK.
|
||
|
|
// Main source: https://www.w3.org/TR/encoding/#gbk-encoder
|
||
|
|
// http://icu-project.org/docs/papers/gb18030.html
|
||
|
|
// http://source.icu-project.org/repos/icu/data/trunk/charset/data/xml/gb-18030-2000.xml
|
||
|
|
// http://www.khngai.com/chinese/charmap/tblgbk.php?page=0
|
||
|
|
'gb18030': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp936.json').concat(require('./tables/gbk-added.json')) },
|
||
|
|
gb18030: function() { return require('./tables/gb18030-ranges.json') },
|
||
|
|
encodeSkipVals: [0x80],
|
||
|
|
encodeAdd: {'€': 0xA2E3},
|
||
|
|
},
|
||
|
|
|
||
|
|
'chinese': 'gb18030',
|
||
|
|
|
||
|
|
|
||
|
|
// == Korean ===============================================================
|
||
|
|
// EUC-KR, KS_C_5601 and KS X 1001 are exactly the same.
|
||
|
|
'windows949': 'cp949',
|
||
|
|
'ms949': 'cp949',
|
||
|
|
'949': 'cp949',
|
||
|
|
'cp949': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp949.json') },
|
||
|
|
},
|
||
|
|
|
||
|
|
'cseuckr': 'cp949',
|
||
|
|
'csksc56011987': 'cp949',
|
||
|
|
'euckr': 'cp949',
|
||
|
|
'isoir149': 'cp949',
|
||
|
|
'korean': 'cp949',
|
||
|
|
'ksc56011987': 'cp949',
|
||
|
|
'ksc56011989': 'cp949',
|
||
|
|
'ksc5601': 'cp949',
|
||
|
|
|
||
|
|
|
||
|
|
// == Big5/Taiwan/Hong Kong ================================================
|
||
|
|
// There are lots of tables for Big5 and cp950. Please see the following links for history:
|
||
|
|
// http://moztw.org/docs/big5/ http://www.haible.de/bruno/charsets/conversion-tables/Big5.html
|
||
|
|
// Variations, in roughly number of defined chars:
|
||
|
|
// * Windows CP 950: Microsoft variant of Big5. Canonical: http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP950.TXT
|
||
|
|
// * Windows CP 951: Microsoft variant of Big5-HKSCS-2001. Seems to be never public. http://me.abelcheung.org/articles/research/what-is-cp951/
|
||
|
|
// * Big5-2003 (Taiwan standard) almost superset of cp950.
|
||
|
|
// * Unicode-at-on (UAO) / Mozilla 1.8. Falling out of use on the Web. Not supported by other browsers.
|
||
|
|
// * Big5-HKSCS (-2001, -2004, -2008). Hong Kong standard.
|
||
|
|
// many unicode code points moved from PUA to Supplementary plane (U+2XXXX) over the years.
|
||
|
|
// Plus, it has 4 combining sequences.
|
||
|
|
// Seems that Mozilla refused to support it for 10 yrs. https://bugzilla.mozilla.org/show_bug.cgi?id=162431 https://bugzilla.mozilla.org/show_bug.cgi?id=310299
|
||
|
|
// because big5-hkscs is the only encoding to include astral characters in non-algorithmic way.
|
||
|
|
// Implementations are not consistent within browsers; sometimes labeled as just big5.
|
||
|
|
// MS Internet Explorer switches from big5 to big5-hkscs when a patch applied.
|
||
|
|
// Great discussion & recap of what's going on https://bugzilla.mozilla.org/show_bug.cgi?id=912470#c31
|
||
|
|
// In the encoder, it might make sense to support encoding old PUA mappings to Big5 bytes seq-s.
|
||
|
|
// Official spec: http://www.ogcio.gov.hk/en/business/tech_promotion/ccli/terms/doc/2003cmp_2008.txt
|
||
|
|
// http://www.ogcio.gov.hk/tc/business/tech_promotion/ccli/terms/doc/hkscs-2008-big5-iso.txt
|
||
|
|
//
|
||
|
|
// Current understanding of how to deal with Big5(-HKSCS) is in the Encoding Standard, http://encoding.spec.whatwg.org/#big5-encoder
|
||
|
|
// Unicode mapping (http://www.unicode.org/Public/MAPPINGS/OBSOLETE/EASTASIA/OTHER/BIG5.TXT) is said to be wrong.
|
||
|
|
|
||
|
|
'windows950': 'cp950',
|
||
|
|
'ms950': 'cp950',
|
||
|
|
'950': 'cp950',
|
||
|
|
'cp950': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp950.json') },
|
||
|
|
},
|
||
|
|
|
||
|
|
// Big5 has many variations and is an extension of cp950. We use Encoding Standard's as a consensus.
|
||
|
|
'big5': 'big5hkscs',
|
||
|
|
'big5hkscs': {
|
||
|
|
type: '_dbcs',
|
||
|
|
table: function() { return require('./tables/cp950.json').concat(require('./tables/big5-added.json')) },
|
||
|
|
encodeSkipVals: [0xa2cc],
|
||
|
|
},
|
||
|
|
|
||
|
|
'cnbig5': 'big5hkscs',
|
||
|
|
'csbig5': 'big5hkscs',
|
||
|
|
'xxbig5': 'big5hkscs',
|
||
|
|
};
|