/** * CP1255 (Windows-Hebrew) encoding module — uniform-format-export (wave-13, spec 180). * * The Israeli Tax Authority (ITA) uniform-structure export requires files * encoded in Windows-1255 (CP1255). The Worker runtime is UTF-16/UTF-8 native * and has no built-in CP1255 support, so we ship a bundled mapping table. * * Rules: * - ASCII 0x00–0x7F: identity mapping * - Hebrew block (U+05D0–U+05EA) + punctuation: mapped per Windows-1255 codepage * - Characters outside CP1255: transliterate where possible, else strip + log * - TextEncoder is NEVER used for final output bytes (it emits UTF-8) */ // ── CP1255 mapping table (Unicode code-point → CP1255 byte) ────────────────── // Only non-identity entries (codepoint !== byte value) are listed. // ASCII 0x00–0x7F map to themselves (handled inline). const CP1255_TABLE: ReadonlyMap = new Map([ // Windows-1255 0x80–0x9F block (printable extensions) [0x20AC, 0x80], // € EURO SIGN [0x201A, 0x82], // ‚ SINGLE LOW-9 QUOTATION MARK [0x0192, 0x83], // ƒ LATIN SMALL LETTER F WITH HOOK [0x201E, 0x84], // „ DOUBLE LOW-9 QUOTATION MARK [0x2026, 0x85], // … HORIZONTAL ELLIPSIS [0x2020, 0x86], // † DAGGER [0x2021, 0x87], // ‡ DOUBLE DAGGER [0x02C6, 0x88], // ˆ MODIFIER LETTER CIRCUMFLEX ACCENT [0x2030, 0x89], // ‰ PER MILLE SIGN [0x2039, 0x8B], // ‹ SINGLE LEFT-POINTING ANGLE QUOTATION MARK [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 [0x02DC, 0x98], // ˜ SMALL TILDE [0x2122, 0x99], // ™ TRADE MARK SIGN [0x203A, 0x9B], // › SINGLE RIGHT-POINTING ANGLE QUOTATION MARK // 0xA0: NO-BREAK SPACE → 0xA0 (same codepoint, identity in 0xA0–0xBF range for CP1255) [0x00A0, 0xA0], // NO-BREAK SPACE [0x00A2, 0xA2], // ¢ CENT SIGN [0x00A3, 0xA3], // £ POUND SIGN [0x00A4, 0xA4], // ¤ CURRENCY SIGN [0x00A5, 0xA5], // ¥ YEN SIGN [0x00A6, 0xA6], // ¦ BROKEN BAR [0x00A7, 0xA7], // § SECTION SIGN [0x00A8, 0xA8], // ¨ DIAERESIS [0x00A9, 0xA9], // © COPYRIGHT SIGN [0x00D7, 0xAA], // × MULTIPLICATION SIGN (Hebrew block: used at 0xAA in CP1255) [0x00AB, 0xAB], // « LEFT-POINTING DOUBLE ANGLE QUOTATION MARK [0x00AC, 0xAC], // ¬ NOT SIGN [0x00AD, 0xAD], // ­ SOFT HYPHEN [0x00AE, 0xAE], // ® REGISTERED SIGN [0x00AF, 0xAF], // ¯ MACRON [0x00B0, 0xB0], // ° DEGREE SIGN [0x00B1, 0xB1], // ± PLUS-MINUS SIGN [0x00B2, 0xB2], // ² SUPERSCRIPT TWO [0x00B3, 0xB3], // ³ SUPERSCRIPT THREE [0x00B4, 0xB4], // ´ ACUTE ACCENT [0x00B5, 0xB5], // µ MICRO SIGN [0x00B6, 0xB6], // ¶ PILCROW SIGN [0x00B7, 0xB7], // · MIDDLE DOT [0x00B8, 0xB8], // ¸ CEDILLA [0x00B9, 0xB9], // ¹ SUPERSCRIPT ONE [0x00F7, 0xBA], // ÷ DIVISION SIGN [0x00BB, 0xBB], // » RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK [0x00BC, 0xBC], // ¼ VULGAR FRACTION ONE QUARTER [0x00BD, 0xBD], // ½ VULGAR FRACTION ONE HALF [0x00BE, 0xBE], // ¾ VULGAR FRACTION THREE QUARTERS [0x00BF, 0xBF], // ¿ INVERTED QUESTION MARK // Nikud (Hebrew points) 0xC0–0xCF [0x05B0, 0xC0], // ְ HEBREW POINT SHEVA [0x05B1, 0xC1], // ֱ HEBREW POINT HATAF SEGOL [0x05B2, 0xC2], // ֲ HEBREW POINT HATAF PATAH [0x05B3, 0xC3], // ֳ HEBREW POINT HATAF QAMATS [0x05B4, 0xC4], // ִ HEBREW POINT HIRIQ [0x05B5, 0xC5], // ֵ HEBREW POINT TSERE [0x05B6, 0xC6], // ֶ HEBREW POINT SEGOL [0x05B7, 0xC7], // ַ HEBREW POINT PATAH [0x05B8, 0xC8], // ָ HEBREW POINT QAMATS [0x05B9, 0xC9], // ֹ HEBREW POINT HOLAM [0x05BB, 0xCB], // ּ HEBREW POINT QUBUTS [0x05BC, 0xCC], // ּ HEBREW POINT DAGESH OR MAPIQ [0x05BD, 0xCD], // ֽ HEBREW POINT METEG [0x05BE, 0xCE], // ־ HEBREW PUNCTUATION MAQAF [0x05BF, 0xCF], // ֿ HEBREW POINT RAFE [0x05C0, 0xD0], // ׀ HEBREW PUNCTUATION PASEQ [0x05C1, 0xD1], // ׁ HEBREW POINT SHIN DOT [0x05C2, 0xD2], // ׂ HEBREW POINT SIN DOT [0x05C3, 0xD3], // ׃ HEBREW PUNCTUATION SOF PASUQ [0x05F0, 0xD4], // װ HEBREW LIGATURE YIDDISH DOUBLE VAV [0x05F1, 0xD5], // ױ HEBREW LIGATURE YIDDISH VAV YOD [0x05F2, 0xD6], // ײ HEBREW LIGATURE YIDDISH DOUBLE YOD [0x05F3, 0xD7], // ׳ HEBREW PUNCTUATION GERESH [0x05F4, 0xD8], // ״ HEBREW PUNCTUATION GERSHAYIM // Hebrew alphabet 0xE0–0xFA [0x05D0, 0xE0], // א HEBREW LETTER ALEF [0x05D1, 0xE1], // ב HEBREW LETTER BET [0x05D2, 0xE2], // ג HEBREW LETTER GIMEL [0x05D3, 0xE3], // ד HEBREW LETTER DALET [0x05D4, 0xE4], // ה HEBREW LETTER HE [0x05D5, 0xE5], // ו HEBREW LETTER VAV [0x05D6, 0xE6], // ז HEBREW LETTER ZAYIN [0x05D7, 0xE7], // ח HEBREW LETTER HET [0x05D8, 0xE8], // ט HEBREW LETTER TET [0x05D9, 0xE9], // י HEBREW LETTER YOD [0x05DA, 0xEA], // ך HEBREW LETTER FINAL KAF [0x05DB, 0xEB], // כ HEBREW LETTER KAF [0x05DC, 0xEC], // ל HEBREW LETTER LAMED [0x05DD, 0xED], // ם HEBREW LETTER FINAL MEM [0x05DE, 0xEE], // מ HEBREW LETTER MEM [0x05DF, 0xEF], // ן HEBREW LETTER FINAL NUN [0x05E0, 0xF0], // נ HEBREW LETTER NUN [0x05E1, 0xF1], // ס HEBREW LETTER SAMEKH [0x05E2, 0xF2], // ע HEBREW LETTER AYIN [0x05E3, 0xF3], // ף HEBREW LETTER FINAL PE [0x05E4, 0xF4], // פ HEBREW LETTER PE [0x05E5, 0xF5], // ץ HEBREW LETTER FINAL TSADI [0x05E6, 0xF6], // צ HEBREW LETTER TSADI [0x05E7, 0xF7], // ק HEBREW LETTER QOF [0x05E8, 0xF8], // ר HEBREW LETTER RESH [0x05E9, 0xF9], // ש HEBREW LETTER SHIN [0x05EA, 0xFA], // ת HEBREW LETTER TAV ]) // Characters with reasonable ASCII transliterations (for ITA fallback) const TRANSLITERATION: ReadonlyMap = new Map([ [0x2018, "'"], // ' → ' [0x2019, "'"], // ' → ' [0x201C, '"'], // " → " [0x201D, '"'], // " → " [0x2013, '-'], // – → - [0x2014, '-'], // — → - [0x2026, '...'], // … → ... [0x00A0, ' '], // NBSP → space ]) /** * Encode a string to CP1255 bytes. * Characters outside CP1255 are stripped (ITA fallback: drop unrepresentable chars). * Use encodeCP1255WithReport to learn which characters were stripped. */ export function encodeCP1255(line: string): Uint8Array { return encodeCP1255WithReport(line).bytes } /** * Encode a string to CP1255 bytes, reporting any characters that were * substituted or stripped. */ export function encodeCP1255WithReport(line: string): { bytes: Uint8Array; substitutions: string[] } { const substitutions: string[] = [] // Pre-estimate: most lines are ASCII, so pre-allocate line.length bytes const buf = new Uint8Array(line.length * 2) // worst case with transliterations let pos = 0 for (let i = 0; i < line.length; i++) { const cp = line.codePointAt(i)! // Advance past surrogate pairs if (cp > 0xFFFF) i++ // ASCII identity range if (cp < 0x80) { buf[pos++] = cp continue } // Check the mapping table const byte = CP1255_TABLE.get(cp) if (byte !== undefined) { buf[pos++] = byte continue } // Check for ASCII transliteration const trans = TRANSLITERATION.get(cp) if (trans !== undefined) { substitutions.push(`U+${cp.toString(16).toUpperCase().padStart(4, '0')} → "${trans}"`) for (let j = 0; j < trans.length; j++) { buf[pos++] = trans.charCodeAt(j) } continue } // Unrepresentable — strip and log const char = String.fromCodePoint(cp) substitutions.push(`U+${cp.toString(16).toUpperCase().padStart(4, '0')} "${char}" stripped`) // byte not written → character dropped } return { bytes: buf.slice(0, pos), substitutions } }