forked from sheetjs/sheetjs
82 lines
3.0 KiB
JavaScript
82 lines
3.0 KiB
JavaScript
var current_codepage = 1200, current_ansi = 1252;
|
|
/*:: declare var cptable:any; */
|
|
/*global cptable:true, window */
|
|
var $cptable;
|
|
|
|
var VALID_ANSI = [ 874, 932, 936, 949, 950, 1250, 1251, 1252, 1253, 1254, 1255, 1256, 1257, 1258, 10000 ];
|
|
/* ECMA-376 Part I 18.4.1 charset to codepage mapping */
|
|
var CS2CP = ({
|
|
0: 1252, /* ANSI */
|
|
1: 65001, /* DEFAULT */
|
|
2: 65001, /* SYMBOL */
|
|
77: 10000, /* MAC */
|
|
128: 932, /* SHIFTJIS */
|
|
129: 949, /* HANGUL */
|
|
130: 1361, /* JOHAB */
|
|
134: 936, /* GB2312 */
|
|
136: 950, /* CHINESEBIG5 */
|
|
161: 1253, /* GREEK */
|
|
162: 1254, /* TURKISH */
|
|
163: 1258, /* VIETNAMESE */
|
|
177: 1255, /* HEBREW */
|
|
178: 1256, /* ARABIC */
|
|
186: 1257, /* BALTIC */
|
|
204: 1251, /* RUSSIAN */
|
|
222: 874, /* THAI */
|
|
238: 1250, /* EASTEUROPE */
|
|
255: 1252, /* OEM */
|
|
69: 6969 /* MISC */
|
|
}/*:any*/);
|
|
|
|
var set_ansi = function(cp/*:number*/) { if(VALID_ANSI.indexOf(cp) == -1) return; current_ansi = CS2CP[0] = cp; };
|
|
function reset_ansi() { set_ansi(1252); }
|
|
|
|
var set_cp = function(cp/*:number*/) { current_codepage = cp; set_ansi(cp); };
|
|
function reset_cp() { set_cp(1200); reset_ansi(); }
|
|
|
|
function char_codes(data/*:string*/)/*:Array<number>*/ { var o/*:Array<number>*/ = []; for(var i = 0, len = data.length; i < len; ++i) o[i] = data.charCodeAt(i); return o; }
|
|
|
|
function utf16leread(data/*:string*/)/*:string*/ {
|
|
var o/*:Array<string>*/ = [];
|
|
for(var i = 0; i < (data.length>>1); ++i) o[i] = String.fromCharCode(data.charCodeAt(2*i) + (data.charCodeAt(2*i+1)<<8));
|
|
return o.join("");
|
|
}
|
|
function utf16lereadu(data/*:Uint8Array*/)/*:string*/ {
|
|
var o/*:Array<string>*/ = [];
|
|
for(var i = 0; i < (data.length>>1); ++i) o[i] = String.fromCharCode(data[2*i] + (data[2*i+1]<<8));
|
|
return o.join("");
|
|
}
|
|
function utf16beread(data/*:string*/)/*:string*/ {
|
|
var o/*:Array<string>*/ = [];
|
|
for(var i = 0; i < (data.length>>1); ++i) o[i] = String.fromCharCode(data.charCodeAt(2*i+1) + (data.charCodeAt(2*i)<<8));
|
|
return o.join("");
|
|
}
|
|
|
|
var debom = function(data/*:string*/)/*:string*/ {
|
|
var c1 = data.charCodeAt(0), c2 = data.charCodeAt(1);
|
|
if(c1 == 0xFF && c2 == 0xFE) return utf16leread(data.slice(2));
|
|
if(c1 == 0xFE && c2 == 0xFF) return utf16beread(data.slice(2));
|
|
if(c1 == 0xFEFF) return data.slice(1);
|
|
return data;
|
|
};
|
|
|
|
var _getchar = function _gc1(x/*:number*/)/*:string*/ { return String.fromCharCode(x); };
|
|
var _getansi = function _ga1(x/*:number*/)/*:string*/ { return String.fromCharCode(x); };
|
|
|
|
function set_cptable(cptable) {
|
|
$cptable = cptable;
|
|
set_cp = function(cp/*:number*/) { current_codepage = cp; set_ansi(cp); };
|
|
debom = function(data/*:string*/) {
|
|
if(data.charCodeAt(0) === 0xFF && data.charCodeAt(1) === 0xFE) { return $cptable.utils.decode(1200, char_codes(data.slice(2))); }
|
|
return data;
|
|
};
|
|
_getchar = function _gc2(x/*:number*/)/*:string*/ {
|
|
if(current_codepage === 1200) return String.fromCharCode(x);
|
|
return $cptable.utils.decode(current_codepage, [x&255,x>>8])[0];
|
|
};
|
|
_getansi = function _ga2(x/*:number*/)/*:string*/ {
|
|
return $cptable.utils.decode(current_ansi, [x])[0];
|
|
};
|
|
cpdoit();
|
|
}
|