Mihaela Mihaljevic
72eaeeb439
String input copies its UTF-8 once into owned storage (an unsafe pointer encapsulated in a class lifetime, never escaping the parser) and scalars decode straight off the pointer, eliminating the per-character closure chain; ASCII literals match and consume at the byte level; and plain-ASCII character data scans as a single byte run with one String built per run. Parse drops 32 percent cumulative (0.49 s to 0.34 s on the 4.2 MB corpus, libxml2 ratio 30x to 20x). The W3C suite caught one boundary (a run starting with '>' could swallow the ]]> check after the character path consumed ]]); the run scanner leaves a leading '>' to the character path, pinned as a unit regression. All conformance suites, Linux, and WASM verified.
27 lines
957 B
Swift
27 lines
957 B
Swift
extension PureXML.Parsing {
|
|
/// Owned contiguous UTF-8 storage for the Reader's fast path: one copy
|
|
/// of the input, read through an unsafe cursor with no per-byte bounds
|
|
/// or retain traffic. The bytes come from a Swift String, so they are
|
|
/// valid UTF-8 by construction. The pointer never escapes the parser;
|
|
/// deallocation rides the class lifetime.
|
|
final class ByteStorage {
|
|
let pointer: UnsafeMutablePointer<UInt8>
|
|
let count: Int
|
|
|
|
init(_ bytes: [UInt8]) {
|
|
count = bytes.count
|
|
pointer = UnsafeMutablePointer<UInt8>.allocate(capacity: max(count, 1))
|
|
bytes.withUnsafeBufferPointer { buffer in
|
|
if let base = buffer.baseAddress {
|
|
pointer.initialize(from: base, count: count)
|
|
}
|
|
}
|
|
}
|
|
|
|
deinit {
|
|
pointer.deinitialize(count: count)
|
|
pointer.deallocate()
|
|
}
|
|
}
|
|
}
|