aboutsummaryrefslogtreecommitdiffstats
path: root/node_modules/iconv-lite/encodings/utf32.js
diff options
context:
space:
mode:
Diffstat (limited to 'node_modules/iconv-lite/encodings/utf32.js')
-rw-r--r--node_modules/iconv-lite/encodings/utf32.js314
1 files changed, 314 insertions, 0 deletions
diff --git a/node_modules/iconv-lite/encodings/utf32.js b/node_modules/iconv-lite/encodings/utf32.js
new file mode 100644
index 0000000..7e49d7a
--- /dev/null
+++ b/node_modules/iconv-lite/encodings/utf32.js
@@ -0,0 +1,314 @@
+"use strict"
+
+var Buffer = require("safer-buffer").Buffer
+
+// == UTF32-LE/BE codec. ==========================================================
+
+exports._utf32 = Utf32Codec
+
+function Utf32Codec (codecOptions, iconv) {
+ this.iconv = iconv
+ this.bomAware = true
+ this.isLE = codecOptions.isLE
+}
+
+exports.utf32le = { type: "_utf32", isLE: true }
+exports.utf32be = { type: "_utf32", isLE: false }
+
+// Aliases
+exports.ucs4le = "utf32le"
+exports.ucs4be = "utf32be"
+
+Utf32Codec.prototype.encoder = Utf32Encoder
+Utf32Codec.prototype.decoder = Utf32Decoder
+
+// -- Encoding
+
+function Utf32Encoder (options, codec) {
+ this.isLE = codec.isLE
+ this.highSurrogate = 0
+}
+
+Utf32Encoder.prototype.write = function (str) {
+ var src = Buffer.from(str, "ucs2")
+ // src.length * 2 covers this chunk's code units (4 bytes each); the extra 4 bytes leave room for a
+ // high surrogate held over from a previous chunk, which is flushed ahead of this chunk's units.
+ var dst = Buffer.alloc(src.length * 2 + 4)
+ var write32 = this.isLE ? dst.writeUInt32LE : dst.writeUInt32BE
+ var offset = 0
+
+ for (var i = 0; i < src.length; i += 2) {
+ var code = src.readUInt16LE(i)
+ var isHighSurrogate = (code >= 0xD800 && code < 0xDC00)
+ var isLowSurrogate = (code >= 0xDC00 && code < 0xE000)
+
+ if (this.highSurrogate) {
+ if (isHighSurrogate || !isLowSurrogate) {
+ // There shouldn't be two high surrogates in a row, nor a high surrogate which isn't followed by a low
+ // surrogate. If this happens, keep the pending high surrogate as a stand-alone semi-invalid character
+ // (technically wrong, but expected by some applications, like Windows file names).
+ write32.call(dst, this.highSurrogate, offset)
+ offset += 4
+ } else {
+ // Create 32-bit value from high and low surrogates;
+ var codepoint = (((this.highSurrogate - 0xD800) << 10) | (code - 0xDC00)) + 0x10000
+
+ write32.call(dst, codepoint, offset)
+ offset += 4
+ this.highSurrogate = 0
+
+ continue
+ }
+ }
+
+ if (isHighSurrogate) { this.highSurrogate = code } else {
+ // Even if the current character is a low surrogate, with no previous high surrogate, we'll
+ // encode it as a semi-invalid stand-alone character for the same reasons expressed above for
+ // unpaired high surrogates.
+ write32.call(dst, code, offset)
+ offset += 4
+ this.highSurrogate = 0
+ }
+ }
+
+ if (offset < dst.length) { dst = dst.slice(0, offset) }
+
+ return dst
+}
+
+Utf32Encoder.prototype.end = function () {
+ // Treat any leftover high surrogate as a semi-valid independent character.
+ if (!this.highSurrogate) { return }
+
+ var buf = Buffer.alloc(4)
+
+ if (this.isLE) { buf.writeUInt32LE(this.highSurrogate, 0) } else { buf.writeUInt32BE(this.highSurrogate, 0) }
+
+ this.highSurrogate = 0
+
+ return buf
+}
+
+// -- Decoding
+
+function Utf32Decoder (options, codec) {
+ this.isLE = codec.isLE
+ this.badChar = codec.iconv.defaultCharUnicode.charCodeAt(0)
+ this.overflow = []
+}
+
+Utf32Decoder.prototype.write = function (src) {
+ if (src.length === 0) { return "" }
+
+ var i = 0
+ var codepoint = 0
+ var dst = Buffer.alloc(src.length + 4)
+ var offset = 0
+ var isLE = this.isLE
+ var overflow = this.overflow
+ var badChar = this.badChar
+
+ if (overflow.length > 0) {
+ for (; i < src.length && overflow.length < 4; i++) { overflow.push(src[i]) }
+
+ if (overflow.length === 4) {
+ // NOTE: codepoint is a signed int32 and can be negative.
+ // NOTE: We copied this block from below to help V8 optimize it (it works with array, not buffer).
+ if (isLE) {
+ codepoint = overflow[0] | (overflow[1] << 8) | (overflow[2] << 16) | (overflow[3] << 24)
+ } else {
+ codepoint = overflow[3] | (overflow[2] << 8) | (overflow[1] << 16) | (overflow[0] << 24)
+ }
+ overflow.length = 0
+
+ offset = _writeCodepoint(dst, offset, codepoint, badChar)
+ }
+ }
+
+ // Main loop. Should be as optimized as possible.
+ for (; i < src.length - 3; i += 4) {
+ // NOTE: codepoint is a signed int32 and can be negative.
+ if (isLE) {
+ codepoint = src[i] | (src[i + 1] << 8) | (src[i + 2] << 16) | (src[i + 3] << 24)
+ } else {
+ codepoint = src[i + 3] | (src[i + 2] << 8) | (src[i + 1] << 16) | (src[i] << 24)
+ }
+ offset = _writeCodepoint(dst, offset, codepoint, badChar)
+ }
+
+ // Keep overflowing bytes.
+ for (; i < src.length; i++) {
+ overflow.push(src[i])
+ }
+
+ return dst.slice(0, offset).toString("ucs2")
+}
+
+function _writeCodepoint (dst, offset, codepoint, badChar) {
+ // NOTE: codepoint is signed int32 and can be negative. We keep it that way to help V8 with optimizations.
+ if (codepoint < 0 || codepoint > 0x10FFFF) {
+ // Not a valid Unicode codepoint
+ codepoint = badChar
+ }
+
+ // Ephemeral Planes: Write high surrogate.
+ if (codepoint >= 0x10000) {
+ codepoint -= 0x10000
+
+ var high = 0xD800 | (codepoint >> 10)
+ dst[offset++] = high & 0xff
+ dst[offset++] = high >> 8
+
+ // Low surrogate is written below.
+ var codepoint = 0xDC00 | (codepoint & 0x3FF)
+ }
+
+ // Write BMP char or low surrogate.
+ dst[offset++] = codepoint & 0xff
+ dst[offset++] = codepoint >> 8
+
+ return offset
+};
+
+Utf32Decoder.prototype.end = function () {
+ if (this.overflow.length === 0) { return }
+
+ // A leftover, incomplete 4-byte code unit at the end of the input is ill-formed. Substitute a
+ // single U+FFFD (Unicode Standard conformance clause C10) instead of silently dropping the bytes.
+ this.overflow.length = 0
+ return String.fromCharCode(this.badChar)
+}
+
+// == UTF-32 Auto codec =============================================================
+// Decoder chooses automatically from UTF-32LE and UTF-32BE using BOM and space-based heuristic.
+// Defaults to UTF-32LE. http://en.wikipedia.org/wiki/UTF-32
+// Encoder/decoder default can be changed: iconv.decode(buf, 'utf32', {defaultEncoding: 'utf-32be'});
+
+// Encoder prepends BOM (which can be overridden with (addBOM: false}).
+
+exports.utf32 = Utf32AutoCodec
+exports.ucs4 = "utf32"
+
+function Utf32AutoCodec (options, iconv) {
+ this.iconv = iconv
+}
+
+Utf32AutoCodec.prototype.encoder = Utf32AutoEncoder
+Utf32AutoCodec.prototype.decoder = Utf32AutoDecoder
+
+// -- Encoding
+
+function Utf32AutoEncoder (options, codec) {
+ options = options || {}
+
+ if (options.addBOM === undefined) {
+ options.addBOM = true
+ }
+
+ this.encoder = codec.iconv.getEncoder(options.defaultEncoding || "utf-32le", options)
+}
+
+Utf32AutoEncoder.prototype.write = function (str) {
+ return this.encoder.write(str)
+}
+
+Utf32AutoEncoder.prototype.end = function () {
+ return this.encoder.end()
+}
+
+// -- Decoding
+
+function Utf32AutoDecoder (options, codec) {
+ this.decoder = null
+ this.initialBufs = []
+ this.initialBufsLen = 0
+ this.options = options || {}
+ this.iconv = codec.iconv
+}
+
+Utf32AutoDecoder.prototype.write = function (buf) {
+ if (!this.decoder) {
+ // Codec is not chosen yet. Accumulate initial bytes.
+ this.initialBufs.push(buf)
+ this.initialBufsLen += buf.length
+
+ if (this.initialBufsLen < 32) // We need more bytes to use space heuristic (see below)
+ { return "" }
+
+ // We have enough bytes -> detect endianness.
+ var encoding = detectEncoding(this.initialBufs, this.options.defaultEncoding)
+ this.decoder = this.iconv.getDecoder(encoding, this.options)
+
+ var resStr = ""
+ for (var i = 0; i < this.initialBufs.length; i++) { resStr += this.decoder.write(this.initialBufs[i]) }
+
+ this.initialBufs.length = this.initialBufsLen = 0
+ return resStr
+ }
+
+ return this.decoder.write(buf)
+}
+
+Utf32AutoDecoder.prototype.end = function () {
+ if (!this.decoder) {
+ var encoding = detectEncoding(this.initialBufs, this.options.defaultEncoding)
+ this.decoder = this.iconv.getDecoder(encoding, this.options)
+
+ var resStr = ""
+ for (var i = 0; i < this.initialBufs.length; i++) { resStr += this.decoder.write(this.initialBufs[i]) }
+
+ var trail = this.decoder.end()
+ if (trail) { resStr += trail }
+
+ this.initialBufs.length = this.initialBufsLen = 0
+ return resStr
+ }
+
+ return this.decoder.end()
+}
+
+function detectEncoding (bufs, defaultEncoding) {
+ var b = []
+ var charsProcessed = 0
+ var invalidLE = 0; var invalidBE = 0 // Number of invalid chars when decoded as LE or BE.
+ var bmpCharsLE = 0; var bmpCharsBE = 0 // Number of BMP chars when decoded as LE or BE.
+
+ outerLoop:
+ for (var i = 0; i < bufs.length; i++) {
+ var buf = bufs[i]
+ for (var j = 0; j < buf.length; j++) {
+ b.push(buf[j])
+ if (b.length === 4) {
+ if (charsProcessed === 0) {
+ // Check BOM first.
+ if (b[0] === 0xFF && b[1] === 0xFE && b[2] === 0 && b[3] === 0) {
+ return "utf-32le"
+ }
+ if (b[0] === 0 && b[1] === 0 && b[2] === 0xFE && b[3] === 0xFF) {
+ return "utf-32be"
+ }
+ }
+
+ if (b[0] !== 0 || b[1] > 0x10) invalidBE++
+ if (b[3] !== 0 || b[2] > 0x10) invalidLE++
+
+ if (b[0] === 0 && b[1] === 0 && (b[2] !== 0 || b[3] !== 0)) bmpCharsBE++
+ if ((b[0] !== 0 || b[1] !== 0) && b[2] === 0 && b[3] === 0) bmpCharsLE++
+
+ b.length = 0
+ charsProcessed++
+
+ if (charsProcessed >= 100) {
+ break outerLoop
+ }
+ }
+ }
+ }
+
+ // Make decisions.
+ if (bmpCharsBE - invalidBE > bmpCharsLE - invalidLE) return "utf-32be"
+ if (bmpCharsBE - invalidBE < bmpCharsLE - invalidLE) return "utf-32le"
+
+ // Couldn't decide (likely all zeros or not enough data).
+ return defaultEncoding || "utf-32le"
+}