feat(emitting): cdataAsText and asciiOnly save options

- cdataAsText emits a CDATA section's content as escaped text instead of
  a <![CDATA[ ... ]]> section.
- asciiOnly escapes every non-ASCII character as a numeric character
  reference, so the output is pure ASCII (libxml2's character-reference
  encoding).

Both apply to the Serializer and the incremental Writer through the
shared Escaping rules.

Closes #91
This commit is contained in:
Mihaela Mihaljevic 2026-06-08 16:34:57 +02:00
parent 85c7bded45
commit e2ba6a7359
5 changed files with 75 additions and 11 deletions

View File

@ -5,20 +5,31 @@ extension PureXML.Emitting {
/// quote and the whitespace characters as numeric references (matching
/// libxml2), so attributes survive attribute-value normalization.
enum Escaping {
static func text(_ value: String) -> String {
static func text(_ value: String, asciiOnly: Bool = false) -> String {
var result = ""
for character in value {
switch character {
case "&": result += "&amp;"
case "<": result += "&lt;"
case ">": result += "&gt;"
default: result.append(character)
default: result += plain(character, asciiOnly: asciiOnly)
}
}
return result
}
static func attribute(_ value: String, quote: Character = "\"") -> String {
/// A character verbatim, or, in ASCII-only mode, each of its non-ASCII
/// scalars as a numeric character reference so the output is pure ASCII.
private static func plain(_ character: Character, asciiOnly: Bool) -> String {
guard asciiOnly else { return String(character) }
var result = ""
for scalar in character.unicodeScalars {
result += scalar.value > 0x7F ? "&#x\(String(scalar.value, radix: 16, uppercase: true));" : String(scalar)
}
return result
}
static func attribute(_ value: String, quote: Character = "\"", asciiOnly: Bool = false) -> String {
var result = ""
for character in value {
if character == quote {
@ -32,7 +43,7 @@ extension PureXML.Emitting {
case "\t": result += "&#9;"
case "\n": result += "&#10;"
case "\r": result += "&#13;"
default: result.append(character)
default: result += plain(character, asciiOnly: asciiOnly)
}
}
return result

View File

@ -24,6 +24,13 @@ public extension PureXML.Emitting {
public var lineEnding: String
/// Whether childless elements collapse to `<name/>`.
public var selfCloseEmptyElements: Bool
/// Whether to emit a CDATA section's content as escaped text rather than as
/// a `<![CDATA[ ... ]]>` section.
public var cdataAsText: Bool
/// Whether to escape every non-ASCII character as a numeric character
/// reference, so the output is pure ASCII (libxml2's character-reference
/// encoding).
public var asciiOnly: Bool
/// Whether to emit an `<?xml ...?>` declaration (libxml2 emits one by
/// default when saving a document; PureXML defaults to off for fragments).
public var includeXMLDeclaration: Bool
@ -40,6 +47,8 @@ public extension PureXML.Emitting {
attributeQuote: QuoteStyle = .double,
lineEnding: String = "\n",
selfCloseEmptyElements: Bool = true,
cdataAsText: Bool = false,
asciiOnly: Bool = false,
includeXMLDeclaration: Bool = false,
xmlVersion: String = "1.0",
encodingName: String? = "UTF-8",
@ -50,6 +59,8 @@ public extension PureXML.Emitting {
self.attributeQuote = attributeQuote
self.lineEnding = lineEnding
self.selfCloseEmptyElements = selfCloseEmptyElements
self.cdataAsText = cdataAsText
self.asciiOnly = asciiOnly
self.includeXMLDeclaration = includeXMLDeclaration
self.xmlVersion = xmlVersion
self.encodingName = encodingName

View File

@ -65,9 +65,11 @@ public extension PureXML.Emitting {
case let .element(element):
write(element, depth: depth, formatted: formatted, output: &output, stack: &stack)
case let .text(value):
output += Escaping.text(value)
output += Escaping.text(value, asciiOnly: options.asciiOnly)
case let .cdata(value):
output += "<![CDATA[\(value)]]>"
output += options.cdataAsText
? Escaping.text(value, asciiOnly: options.asciiOnly)
: "<![CDATA[\(value)]]>"
case let .comment(value):
output += "<!--\(value)-->"
case let .processingInstruction(target, data):
@ -88,7 +90,8 @@ public extension PureXML.Emitting {
output += "<" + element.name.description
let quote = options.attributeQuote.character
for attribute in element.attributes {
output += " \(attribute.name.description)=\(quote)\(Escaping.attribute(attribute.value, quote: quote))\(quote)"
let value = Escaping.attribute(attribute.value, quote: quote, asciiOnly: options.asciiOnly)
output += " \(attribute.name.description)=\(quote)\(value)\(quote)"
}
if element.children.isEmpty, options.selfCloseEmptyElements {

View File

@ -71,20 +71,21 @@ public extension PureXML.Emitting {
public mutating func writeAttribute(_ name: String, _ value: String) {
guard startTagOpen else { return }
let quote = options.attributeQuote.character
output += " \(name)=\(quote)\(Escaping.attribute(value, quote: quote))\(quote)"
output += " \(name)=\(quote)\(Escaping.attribute(value, quote: quote, asciiOnly: options.asciiOnly))\(quote)"
}
/// Writes escaped character data into the current element.
public mutating func writeString(_ text: String) {
closeStartTag()
output += Escaping.text(text)
output += Escaping.text(text, asciiOnly: options.asciiOnly)
setTopContent(.text)
}
/// Writes a `<![CDATA[ ... ]]>` section.
/// Writes a `<![CDATA[ ... ]]>` section, or its escaped text when
/// `cdataAsText` is set.
public mutating func writeCData(_ text: String) {
closeStartTag()
output += "<![CDATA[\(text)]]>"
output += options.cdataAsText ? Escaping.text(text, asciiOnly: options.asciiOnly) : "<![CDATA[\(text)]]>"
setTopContent(.text)
}

View File

@ -0,0 +1,38 @@
@testable import PureXML
import Testing
@Suite("Serialization: CDATA-as-text and ASCII-only")
struct SerializationEscapeTests {
private func serialize(_ node: PureXML.Model.Node, _ options: PureXML.Emitting.Options) -> String {
PureXML.serialize(node, options: options)
}
@Test("cdataAsText emits a CDATA section as escaped text")
func test_cdataAsText() {
let element = PureXML.Model.Element("a", children: [.cdata("x < y & z")])
let asSection = serialize(.element(element), PureXML.Emitting.Options(prettyPrint: false))
#expect(asSection == "<a><![CDATA[x < y & z]]></a>")
let asText = serialize(.element(element), PureXML.Emitting.Options(prettyPrint: false, cdataAsText: true))
#expect(asText == "<a>x &lt; y &amp; z</a>")
}
@Test("asciiOnly escapes non-ASCII characters as numeric references")
func test_asciiOnlyText() {
let element = PureXML.Model.Element("a", children: [.text("café")])
let options = PureXML.Emitting.Options(prettyPrint: false, asciiOnly: true)
#expect(serialize(.element(element), options) == "<a>caf&#xE9;</a>")
}
@Test("asciiOnly escapes non-ASCII in attribute values too")
func test_asciiOnlyAttribute() {
let element = PureXML.Model.Element("a", attributes: [.init("t", "ñ")])
let options = PureXML.Emitting.Options(prettyPrint: false, asciiOnly: true)
#expect(serialize(.element(element), options) == "<a t=\"&#xF1;\"/>")
}
@Test("Without asciiOnly, non-ASCII is left verbatim")
func test_nonAsciiVerbatim() {
let element = PureXML.Model.Element("a", children: [.text("café")])
#expect(serialize(.element(element), PureXML.Emitting.Options(prettyPrint: false)) == "<a>café</a>")
}
}