Index: src/encodings/codepage_850.m ================================================================== --- src/encodings/codepage_850.m +++ src/encodings/codepage_850.m @@ -15,10 +15,12 @@ */ #include "config.h" #import "OFString.h" + +#import "common.h" const of_char16_t of_codepage_850_table[] = { 0x00C7, 0x00FC, 0x00E9, 0x00E2, 0x00E4, 0x00E0, 0x00E5, 0x00E7, 0x00EA, 0x00EB, 0x00E8, 0x00EF, 0x00EE, 0x00EC, 0x00C4, 0x00C5, 0x00C9, 0x00E6, 0x00C6, 0x00F4, 0x00F6, 0x00F2, 0x00FB, 0x00F9, @@ -37,420 +39,109 @@ 0x00B0, 0x00A8, 0x00B7, 0x00B9, 0x00B3, 0x00B2, 0x25A0, 0x00A0 }; const size_t of_codepage_850_table_offset = 256 - (sizeof(of_codepage_850_table) / sizeof(*of_codepage_850_table)); + +static const char page0[] = { + 0xFF, 0xAD, 0xBD, 0x9C, 0xCF, 0xBE, 0xDD, 0xF5, + 0xF9, 0xB8, 0xA6, 0xAE, 0xAA, 0xF0, 0xA9, 0xEE, + 0xF8, 0xF1, 0xFD, 0xFC, 0xEF, 0xE6, 0xF4, 0xFA, + 0xF7, 0xFB, 0xA7, 0xAF, 0xAC, 0xAB, 0xF3, 0xA8, + 0xB7, 0xB5, 0xB6, 0xC7, 0x8E, 0x8F, 0x92, 0x80, + 0xD4, 0x90, 0xD2, 0xD3, 0xDE, 0xD6, 0xD7, 0xD8, + 0xD1, 0xA5, 0xE3, 0xE0, 0xE2, 0xE5, 0x99, 0x9E, + 0x9D, 0xEB, 0xE9, 0xEA, 0x9A, 0xED, 0xE8, 0xE1, + 0x85, 0xA0, 0x83, 0xC6, 0x84, 0x86, 0x91, 0x87, + 0x8A, 0x82, 0x88, 0x89, 0x8D, 0xA1, 0x8C, 0x8B, + 0xD0, 0xA4, 0x95, 0xA2, 0x93, 0xE4, 0x94, 0xF6, + 0x9B, 0x97, 0xA3, 0x96, 0x81, 0xEC, 0xE7, 0x98 +}; +static const uint8_t page0Start = 0xA0; +static const uint16_t page0Size = sizeof(page0) / sizeof(*page0); + +static const char page1[] = { + 0xD5, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x9F +}; +static const uint8_t page1Start = 0x31; +static const uint16_t page1Size = sizeof(page1) / sizeof(*page1); + +static const char page20[] = { + 0xF2 +}; +static const uint8_t page20Start = 0x17; +static const uint16_t page20Size = sizeof(page20) / sizeof(*page20); + +static const char page25[] = { + 0xC4, 0x00, 0xB3, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xDA, 0x00, 0x00, 0x00, + 0xBF, 0x00, 0x00, 0x00, 0xC0, 0x00, 0x00, 0x00, + 0xD9, 0x00, 0x00, 0x00, 0xC3, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xB4, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xC2, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xC1, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xC5, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xCD, 0xBA, 0x00, 0x00, 0xC9, 0x00, 0x00, 0xBB, + 0x00, 0x00, 0xC8, 0x00, 0x00, 0xBC, 0x00, 0x00, + 0xCC, 0x00, 0x00, 0xB9, 0x00, 0x00, 0xCB, 0x00, + 0x00, 0xCA, 0x00, 0x00, 0xCE, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xDF, 0x00, 0x00, 0x00, 0xDC, 0x00, 0x00, 0x00, + 0xDB, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xB0, 0xB1, 0xB2, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xFE +}; +static const uint8_t page25Start = 0x00; +static const uint16_t page25Size = sizeof(page25) / sizeof(*page25); + bool of_unicode_to_codepage_850(const of_unichar_t *input, unsigned char *output, size_t length, bool lossy) { for (size_t i = 0; i < length; i++) { of_unichar_t c = input[i]; if OF_UNLIKELY (c > 0x7F) { + uint8_t index; + if OF_UNLIKELY (c > 0xFFFF) { if (lossy) { output[i] = '?'; continue; } else return false; } - switch ((of_char16_t)c) { - case 0xC7: - output[i] = 0x80; - break; - case 0xFC: - output[i] = 0x81; - break; - case 0xE9: - output[i] = 0x82; - break; - case 0xE2: - output[i] = 0x83; - break; - case 0xE4: - output[i] = 0x84; - break; - case 0xE0: - output[i] = 0x85; - break; - case 0xE5: - output[i] = 0x86; - break; - case 0xE7: - output[i] = 0x87; - break; - case 0xEA: - output[i] = 0x88; - break; - case 0xEB: - output[i] = 0x89; - break; - case 0xE8: - output[i] = 0x8A; - break; - case 0xEF: - output[i] = 0x8B; - break; - case 0xEE: - output[i] = 0x8C; - break; - case 0xEC: - output[i] = 0x8D; - break; - case 0xC4: - output[i] = 0x8E; - break; - case 0xC5: - output[i] = 0x8F; - break; - case 0xC9: - output[i] = 0x90; - break; - case 0xE6: - output[i] = 0x91; - break; - case 0xC6: - output[i] = 0x92; - break; - case 0xF4: - output[i] = 0x93; - break; - case 0xF6: - output[i] = 0x94; - break; - case 0xF2: - output[i] = 0x95; - break; - case 0xFB: - output[i] = 0x96; - break; - case 0xF9: - output[i] = 0x97; - break; - case 0xFF: - output[i] = 0x98; - break; - case 0xD6: - output[i] = 0x99; - break; - case 0xDC: - output[i] = 0x9A; - break; - case 0xF8: - output[i] = 0x9B; - break; - case 0xA3: - output[i] = 0x9C; - break; - case 0xD8: - output[i] = 0x9D; - break; - case 0xD7: - output[i] = 0x9E; - break; - case 0x192: - output[i] = 0x9F; - break; - case 0xE1: - output[i] = 0xA0; - break; - case 0xED: - output[i] = 0xA1; - break; - case 0xF3: - output[i] = 0xA2; - break; - case 0xFA: - output[i] = 0xA3; - break; - case 0xF1: - output[i] = 0xA4; - break; - case 0xD1: - output[i] = 0xA5; - break; - case 0xAA: - output[i] = 0xA6; - break; - case 0xBA: - output[i] = 0xA7; - break; - case 0xBF: - output[i] = 0xA8; - break; - case 0xAE: - output[i] = 0xA9; - break; - case 0xAC: - output[i] = 0xAA; - break; - case 0xBD: - output[i] = 0xAB; - break; - case 0xBC: - output[i] = 0xAC; - break; - case 0xA1: - output[i] = 0xAD; - break; - case 0xAB: - output[i] = 0xAE; - break; - case 0xBB: - output[i] = 0xAF; - break; - case 0x2591: - output[i] = 0xB0; - break; - case 0x2592: - output[i] = 0xB1; - break; - case 0x2593: - output[i] = 0xB2; - break; - case 0x2502: - output[i] = 0xB3; - break; - case 0x2524: - output[i] = 0xB4; - break; - case 0xC1: - output[i] = 0xB5; - break; - case 0xC2: - output[i] = 0xB6; - break; - case 0xC0: - output[i] = 0xB7; - break; - case 0xA9: - output[i] = 0xB8; - break; - case 0x2563: - output[i] = 0xB9; - break; - case 0x2551: - output[i] = 0xBA; - break; - case 0x2557: - output[i] = 0xBB; - break; - case 0x255D: - output[i] = 0xBC; - break; - case 0xA2: - output[i] = 0xBD; - break; - case 0xA5: - output[i] = 0xBE; - break; - case 0x2510: - output[i] = 0xBF; - break; - case 0x2514: - output[i] = 0xC0; - break; - case 0x2534: - output[i] = 0xC1; - break; - case 0x252C: - output[i] = 0xC2; - break; - case 0x251C: - output[i] = 0xC3; - break; - case 0x2500: - output[i] = 0xC4; - break; - case 0x253C: - output[i] = 0xC5; - break; - case 0xE3: - output[i] = 0xC6; - break; - case 0xC3: - output[i] = 0xC7; - break; - case 0x255A: - output[i] = 0xC8; - break; - case 0x2554: - output[i] = 0xC9; - break; - case 0x2569: - output[i] = 0xCA; - break; - case 0x2566: - output[i] = 0xCB; - break; - case 0x2560: - output[i] = 0xCC; - break; - case 0x2550: - output[i] = 0xCD; - break; - case 0x256C: - output[i] = 0xCE; - break; - case 0xA4: - output[i] = 0xCF; - break; - case 0xF0: - output[i] = 0xD0; - break; - case 0xD0: - output[i] = 0xD1; - break; - case 0xCA: - output[i] = 0xD2; - break; - case 0xCB: - output[i] = 0xD3; - break; - case 0xC8: - output[i] = 0xD4; - break; - case 0x131: - output[i] = 0xD5; - break; - case 0xCD: - output[i] = 0xD6; - break; - case 0xCE: - output[i] = 0xD7; - break; - case 0xCF: - output[i] = 0xD8; - break; - case 0x2518: - output[i] = 0xD9; - break; - case 0x250C: - output[i] = 0xDA; - break; - case 0x2588: - output[i] = 0xDB; - break; - case 0x2584: - output[i] = 0xDC; - break; - case 0xA6: - output[i] = 0xDD; - break; - case 0xCC: - output[i] = 0xDE; - break; - case 0x2580: - output[i] = 0xDF; - break; - case 0xD3: - output[i] = 0xE0; - break; - case 0xDF: - output[i] = 0xE1; - break; - case 0xD4: - output[i] = 0xE2; - break; - case 0xD2: - output[i] = 0xE3; - break; - case 0xF5: - output[i] = 0xE4; - break; - case 0xD5: - output[i] = 0xE5; - break; - case 0xB5: - output[i] = 0xE6; - break; - case 0xFE: - output[i] = 0xE7; - break; - case 0xDE: - output[i] = 0xE8; - break; - case 0xDA: - output[i] = 0xE9; - break; - case 0xDB: - output[i] = 0xEA; - break; - case 0xD9: - output[i] = 0xEB; - break; - case 0xFD: - output[i] = 0xEC; - break; - case 0xDD: - output[i] = 0xED; - break; - case 0xAF: - output[i] = 0xEE; - break; - case 0xB4: - output[i] = 0xEF; - break; - case 0xAD: - output[i] = 0xF0; - break; - case 0xB1: - output[i] = 0xF1; - break; - case 0x2017: - output[i] = 0xF2; - break; - case 0xBE: - output[i] = 0xF3; - break; - case 0xB6: - output[i] = 0xF4; - break; - case 0xA7: - output[i] = 0xF5; - break; - case 0xF7: - output[i] = 0xF6; - break; - case 0xB8: - output[i] = 0xF7; - break; - case 0xB0: - output[i] = 0xF8; - break; - case 0xA8: - output[i] = 0xF9; - break; - case 0xB7: - output[i] = 0xFA; - break; - case 0xB9: - output[i] = 0xFB; - break; - case 0xB3: - output[i] = 0xFC; - break; - case 0xB2: - output[i] = 0xFD; - break; - case 0x25A0: - output[i] = 0xFE; - break; - case 0xA0: - output[i] = 0xFF; - break; + switch (c >> 8) { + CASE_MISSING_IS_ERROR(0) + CASE_MISSING_IS_ERROR(1) + CASE_MISSING_IS_ERROR(20) + CASE_MISSING_IS_ERROR(25) default: - if (lossy) + if (lossy) { output[i] = '?'; - else + continue; + } else return false; - - break; } } else output[i] = (unsigned char)c; } return true; }