PureXML/Sources/Parsing/StreamingDecoder.swift
Mihaela Mihaljevic 204dd3895e feat(parser): streaming byte decode (incremental UTF-8/UTF-16)
Add PureXML.parse(pullingBytes:) and events(pullingBytes:) over an incremental
byte source (() -> UInt8?). StreamingDecoder detects the encoding from the
leading bytes, strips any BOM, and yields one scalar at a time as a Character,
so a byte stream is never fully buffered. UTF-8 (1 to 4 bytes) and UTF-16 (with
surrogate pairs) are decoded incrementally; invalid sequences become the Unicode
replacement character. Standard library only, no Foundation. Reimplemented from
the documented UTF decoding rules; no upstream code copied.
2026-06-08 01:07:24 +02:00

122 lines
4.4 KiB
Swift

extension PureXML.Parsing {
/// Decodes a byte stream into characters incrementally. It detects the
/// encoding from the leading bytes (buffering only those few), strips any
/// byte-order mark, then yields one Unicode scalar at a time as a Character,
/// so a byte source is never fully materialized. Invalid sequences yield the
/// Unicode replacement character rather than failing the stream.
///
/// This works at scalar granularity; a multi-scalar grapheme cluster is
/// surfaced as separate characters, which is harmless for XML tokenization.
struct StreamingDecoder {
private let pull: () -> UInt8?
private var encoding: InputEncoding?
private var prefix: [UInt8] = []
private var prefixIndex = 0
private var finished = false
private static let replacement: Character = "\u{FFFD}"
init(pullingBytes pull: @escaping () -> UInt8?) {
self.pull = pull
}
/// Returns the next decoded character, or nil at the end of the stream.
mutating func next() -> Character? {
if encoding == nil {
detect()
}
guard !finished, let encoding else {
return nil
}
switch encoding {
case .utf8:
return nextUTF8()
case .utf16BigEndian:
return nextUTF16(bigEndian: true)
case .utf16LittleEndian:
return nextUTF16(bigEndian: false)
}
}
private mutating func detect() {
while prefix.count < 4, let byte = pull() {
prefix.append(byte)
}
let detected = InputEncoding.detectWithBOM(prefix)
encoding = detected.encoding
prefixIndex = detected.bomLength
}
private mutating func nextByte() -> UInt8? {
if prefixIndex < prefix.count {
let byte = prefix[prefixIndex]
prefixIndex += 1
return byte
}
return pull()
}
private mutating func nextUTF8() -> Character? {
guard let first = nextByte() else {
finished = true
return nil
}
if first < 0x80 {
return Character(Unicode.Scalar(first))
}
let length: Int
var value: UInt32
if first & 0xE0 == 0xC0 {
length = 2
value = UInt32(first & 0x1F)
} else if first & 0xF0 == 0xE0 {
length = 3
value = UInt32(first & 0x0F)
} else if first & 0xF8 == 0xF0 {
length = 4
value = UInt32(first & 0x07)
} else {
return Self.replacement
}
for _ in 1 ..< length {
guard let continuation = nextByte(), continuation & 0xC0 == 0x80 else {
return Self.replacement
}
value = (value << 6) | UInt32(continuation & 0x3F)
}
guard let scalar = Unicode.Scalar(value) else {
return Self.replacement
}
return Character(scalar)
}
private mutating func nextUTF16(bigEndian: Bool) -> Character? {
guard let unit = nextUnit(bigEndian: bigEndian) else {
finished = true
return nil
}
if (0xD800 ... 0xDBFF).contains(unit) {
guard let low = nextUnit(bigEndian: bigEndian), (0xDC00 ... 0xDFFF).contains(low) else {
return Self.replacement
}
let value = 0x10000 + (UInt32(unit - 0xD800) << 10) + UInt32(low - 0xDC00)
guard let scalar = Unicode.Scalar(value) else {
return Self.replacement
}
return Character(scalar)
}
guard let scalar = Unicode.Scalar(unit) else {
return Self.replacement
}
return Character(scalar)
}
private mutating func nextUnit(bigEndian: Bool) -> UInt16? {
guard let high = nextByte(), let low = nextByte() else {
return nil
}
return bigEndian ? (UInt16(high) << 8 | UInt16(low)) : (UInt16(low) << 8 | UInt16(high))
}
}
}