Commit `845710a22b`

845710a22b57469c87b66752c230ed8c274b4bf3

Verified · cmc

cmc <hello@cleberg.net> · 2026-10-04 23:14 UTC

Add OrgCore package and SourceText byte contract

Layout: unified · split

Package.swift added +14

		@@ -0,0 +1,14 @@
	1	// swift-tools-version: 6.2
	2	import PackageDescription
	3
	4	let package = Package(
	5	name: "Orgstar",
	6	platforms: [.macOS(.v26), .iOS(.v26)],
	7	products: [
	8	.library(name: "OrgCore", targets: ["OrgCore"])
	9	],
	10	targets: [
	11	.target(name: "OrgCore"),
	12	.testTarget(name: "OrgCoreTests", dependencies: ["OrgCore"])
	13	]
	14	)

Sources/OrgCore/SourceText.swift added +40

		@@ -0,0 +1,40 @@
	1	import Foundation
	2
	3	/// A file's bytes and their decoded text. Only UTF-8 (with or without a BOM) is editable;
	4	/// anything else decodes with replacement characters for display and is never written back.
	5	public struct SourceText: Sendable {
	6	public let originalBytes: [UInt8]
	7	public let hasBOM: Bool
	8	public let isValidUTF8: Bool
	9	/// Decoded text without the BOM.
	10	public let text: String
	11
	12	private static let bom: [UInt8] = [0xEF, 0xBB, 0xBF]
	13
	14	public init(bytes: [UInt8]) {
	15	originalBytes = bytes
	16	hasBOM = bytes.starts(with: Self.bom)
	17	let body = hasBOM ? Array(bytes.dropFirst(3)) : bytes
	18	if let decoded = String(validating: body, as: UTF8.self) {
	19	text = decoded
	20	isValidUTF8 = true
	21	} else {
	22	text = String(decoding: body, as: UTF8.self)
	23	isValidUTF8 = false
	24	}
	25	}
	26
	27	public init(_ text: String) {
	28	self.init(bytes: Array(text.utf8))
	29	}
	30
	31	public var isEditable: Bool { isValidUTF8 }
	32
	33	/// Bytes to write for `newText`. Unchanged text returns the original bytes. Valid UTF-8
	34	/// round-trips through `String` unchanged, so untouched spans keep their exact bytes.
	35	public func encode(_ newText: String) -> [UInt8] {
	36	if newText == text { return originalBytes }
	37	precondition(isEditable, "a read-only document cannot be re-encoded")
	38	return (hasBOM ? Self.bom : []) + Array(newText.utf8)
	39	}
	40	}

Tests/OrgCoreTests/SourceTextTests.swift added +40

		@@ -0,0 +1,40 @@
	1	import Testing
	2	@testable import OrgCore
	3
	4	struct SourceTextTests {
	5	@Test func plainUTF8() {
	6	let source = SourceText(bytes: Array("* a\n".utf8))
	7	#expect(source.text == "* a\n")
	8	#expect(!source.hasBOM)
	9	#expect(source.isEditable)
	10	}
	11
	12	@Test func bomIsStrippedAndRestored() {
	13	let bytes: [UInt8] = [0xEF, 0xBB, 0xBF] + Array("x\n".utf8)
	14	let source = SourceText(bytes: bytes)
	15	#expect(source.hasBOM)
	16	#expect(source.text == "x\n")
	17	#expect(source.encode(source.text) == bytes)
	18	#expect(source.encode("y\n") == [0xEF, 0xBB, 0xBF] + Array("y\n".utf8))
	19	}
	20
	21	@Test func crlfAndMixedEndingsSurvive() {
	22	let bytes = Array("a\r\nb\nc\r\n".utf8)
	23	let source = SourceText(bytes: bytes)
	24	#expect(source.encode(source.text) == bytes)
	25	}
	26
	27	@Test func invalidUTF8IsReadOnlyAndUnchanged() {
	28	let bytes: [UInt8] = [0x61, 0xFF, 0x0A]
	29	let source = SourceText(bytes: bytes)
	30	#expect(!source.isValidUTF8)
	31	#expect(!source.isEditable)
	32	#expect(source.encode(source.text) == bytes)
	33	}
	34
	35	@Test func nonBMPAndCombiningSurvive() {
	36	let bytes = Array("😀 e\u{301}\n".utf8)
	37	let source = SourceText(bytes: bytes)
	38	#expect(source.encode(source.text) == bytes)
	39	}
	40	}

Commit 845710a22b