Created
January 9, 2016 08:40
-
-
Save ngrebenshikov/e82ffca794554c67a9af to your computer and use it in GitHub Desktop.
UTF8 to CP1251 string converter on Haxe
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
package ; | |
import haxe.io.Bytes; | |
import haxe.ds.IntMap; | |
class TextEncodingHelper { | |
// Based on https://code.google.com/p/convert-utf8-to-cp1251/source/browse/main.cpp | |
private static var utf8ToCp1251: IntMap<Int> = [ | |
0x201A => 0x82, // SINGLE LOW-9 QUOTATION MARK | |
0x0453 => 0x83, // CYRILLIC SMALL LETTER GJE | |
0x201E => 0x84, // DOUBLE LOW-9 QUOTATION MARK | |
0x2026 => 0x85, // HORIZONTAL ELLIPSIS | |
0x2020 => 0x86, // DAGGER | |
0x2021 => 0x87, // DOUBLE DAGGER | |
0x20AC => 0x88, // EURO SIGN | |
0x2030 => 0x89, // PER MILLE SIGN | |
0x0409 => 0x8A, // CYRILLIC CAPITAL LETTER LJE | |
0x2039 => 0x8B, // SINGLE LEFT-POINTING ANGLE QUOTATION MARK | |
0x040A => 0x8C, // CYRILLIC CAPITAL LETTER NJE | |
0x040C => 0x8D, // CYRILLIC CAPITAL LETTER KJE | |
0x040B => 0x8E, // CYRILLIC CAPITAL LETTER TSHE | |
0x040F => 0x8F, // CYRILLIC CAPITAL LETTER DZHE | |
0x0452 => 0x90, // CYRILLIC SMALL LETTER DJE | |
0x2018 => 0x91, // LEFT SINGLE QUOTATION MARK | |
0x2019 => 0x92, // RIGHT SINGLE QUOTATION MARK | |
0x201C => 0x93, // LEFT DOUBLE QUOTATION MARK | |
0x201D => 0x94, // RIGHT DOUBLE QUOTATION MARK | |
0x2022 => 0x95, // BULLET | |
0x2013 => 0x96, // EN DASH | |
0x2014 => 0x97, // EM DASH | |
0x2122 => 0x99, // TRADE MARK SIGN | |
0x0459 => 0x9A, // CYRILLIC SMALL LETTER LJE | |
0x203A => 0x9B, // SINGLE RIGHT-POINTING ANGLE QUOTATION MARK | |
0x045A => 0x9C, // CYRILLIC SMALL LETTER NJE | |
0x045C => 0x9D, // CYRILLIC SMALL LETTER KJE | |
0x045B => 0x9E, // CYRILLIC SMALL LETTER TSHE | |
0x045F => 0x9F, // CYRILLIC SMALL LETTER DZHE | |
0x00A0 => 0xA0, // NO-BREAK SPACE | |
0x040E => 0xA1, // CYRILLIC CAPITAL LETTER SHORT U | |
0x045E => 0xA2, // CYRILLIC SMALL LETTER SHORT U | |
0x0408 => 0xA3, // CYRILLIC CAPITAL LETTER JE | |
0x00A4 => 0xA4, // CURRENCY SIGN | |
0x0490 => 0xA5, // CYRILLIC CAPITAL LETTER GHE WITH UPTURN | |
0x00A6 => 0xA6, // BROKEN BAR | |
0x00A7 => 0xA7, // SECTION SIGN | |
0x0401 => 0xA8, // CYRILLIC CAPITAL LETTER IO | |
0x00A9 => 0xA9, // COPYRIGHT SIGN | |
0x0404 => 0xAA, // CYRILLIC CAPITAL LETTER UKRAINIAN IE | |
0x00AB => 0xAB, // LEFT-POINTING DOUBLE ANGLE QUOTATION MARK | |
0x00AC => 0xAC, // NOT SIGN | |
0x00AD => 0xAD, // SOFT HYPHEN | |
0x00AE => 0xAE, // REGISTERED SIGN | |
0x0407 => 0xAF, // CYRILLIC CAPITAL LETTER YI | |
0x00B0 => 0xB0, // DEGREE SIGN | |
0x00B1 => 0xB1, // PLUS-MINUS SIGN | |
0x0406 => 0xB2, // CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I | |
0x0456 => 0xB3, // CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I | |
0x0491 => 0xB4, // CYRILLIC SMALL LETTER GHE WITH UPTURN | |
0x00B5 => 0xB5, // MICRO SIGN | |
0x00B6 => 0xB6, // PILCROW SIGN | |
0x00B7 => 0xB7, // MIDDLE DOT | |
0x0451 => 0xB8, // CYRILLIC SMALL LETTER IO | |
0x2116 => 0xB9, // NUMERO SIGN | |
0x0454 => 0xBA, // CYRILLIC SMALL LETTER UKRAINIAN IE | |
0x00BB => 0xBB, // RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK | |
0x0458 => 0xBC, // CYRILLIC SMALL LETTER JE | |
0x0405 => 0xBD, // CYRILLIC CAPITAL LETTER DZE | |
0x0455 => 0xBE, // CYRILLIC SMALL LETTER DZE | |
0x0457 => 0xBF // CYRILLIC SMALL LETTER YI | |
]; | |
private static function convertUtf8ToWindows1251(utf8String: String): String { | |
var utf8: Bytes = Bytes.ofString(utf8String); | |
var result: StringBuf = new StringBuf(); | |
var i: Int = 0; | |
while(i < utf8.length) { | |
var prefix: Int = utf8.get(i); | |
var suffix: Int = utf8.get(i+1); | |
if ((prefix & 0x80) == 0) { | |
result.addChar(prefix); | |
} else if ((~prefix) & 0x20 != 0) { | |
var first5bit: Int = prefix & 0x1F; | |
first5bit <<= 6; | |
var sec6bit: Int = suffix & 0x3F; | |
var unicodeChar: Int = first5bit + sec6bit; | |
if ( unicodeChar >= 0x410 && unicodeChar <= 0x44F ) { | |
result.addChar(unicodeChar - 0x350); | |
} else if (unicodeChar >= 0x80 && unicodeChar <= 0xFF) { | |
result.addChar(unicodeChar); | |
} else if (unicodeChar >= 0x402 && unicodeChar <= 0x403) { | |
result.addChar(unicodeChar - 0x382); | |
} else { | |
if (utf8ToCp1251.exists(unicodeChar)) { | |
result.addChar(utf8ToCp1251.get(unicodeChar)); | |
} else { | |
trace("Can't convert the char"); | |
} | |
} | |
i += 1; | |
} else { | |
trace("Can't convert the char"); | |
} | |
i += 1; | |
} | |
return result.toString(); | |
} | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment