Mihaela Mihaljevic
83f96d68a0
Lock the strict and recovering parse paths together: recovery synthesizes events through machinery the strict path never exercises, so a 21-document corpus of well-formed input (the C14N conformance inputs plus nesting, mixed content, attribute quoting/escapes, namespaces and redeclaration, entities, character references, prolog, comments, CDATA, xml:space) asserts that PureXML.read produces the identical Node tree as PureXML.parse with no diagnostics, and that readTree's editor tree projects to the same node. Both paths agree on every document; the differential now guards them. Closes #116
62 lines
2.7 KiB
Swift
62 lines
2.7 KiB
Swift
@testable import PureXML
|
|
import Testing
|
|
|
|
/// Locks the strict and recovering parse paths together: the recovery code
|
|
/// synthesizes events through machinery the strict path never exercises, so
|
|
/// every well-formed document must produce the identical tree under both modes
|
|
/// with no diagnostics. (#116: a recovery-synthesis bug is invisible to
|
|
/// strict-path tests without this differential.)
|
|
@Suite("Strict vs recovering parser differential")
|
|
struct StrictRecoveringDifferentialTests {
|
|
/// Well-formed documents spanning the constructs both paths must agree on,
|
|
/// including the conformance-corpus inputs (kept in sync by inclusion).
|
|
private let corpus = [
|
|
// The C14N conformance-corpus inputs.
|
|
"<e b=\"2\" a=\"1\"></e>",
|
|
"<e/>",
|
|
"<e><!--c-->x</e>",
|
|
"<e><![CDATA[<x>&]]></e>",
|
|
"<e><?pi data?></e>",
|
|
"<e xmlns=\"urn:x\"></e>",
|
|
// Structure: nesting, siblings, mixed content, whitespace.
|
|
"<r><a><b><c>deep</c></b></a></r>",
|
|
"<r><i>1</i><i>2</i><i>3</i></r>",
|
|
"<p>text <b>bold</b> tail</p>",
|
|
"<r>\n <a/>\n <b/>\n</r>",
|
|
// Attributes: quoting, escapes, namespaces on attributes.
|
|
"<e a=\"x&y\" b=\"<tag>\"/>",
|
|
"<e a='single' b=\"double\"/>",
|
|
"<r xmlns:p=\"urn:x\"><p:c p:attr=\"v\"/></r>",
|
|
// Entities and character references.
|
|
"<e>&<>"'</e>",
|
|
"<e>AB</e>",
|
|
// Prolog and misc constructs.
|
|
"<?xml version=\"1.0\"?><r/>",
|
|
"<!-- leading --><r/><!-- trailing -->",
|
|
"<r><![CDATA[]]></r>",
|
|
"<r xml:space=\"preserve\"> kept </r>",
|
|
// Default and redeclared namespaces.
|
|
"<r xmlns=\"urn:a\"><c xmlns=\"urn:b\"><d/></c></r>",
|
|
]
|
|
|
|
@Test("Well-formed documents parse identically in strict and recovering modes")
|
|
func test_treesAgree() throws {
|
|
for document in corpus {
|
|
let strict = try PureXML.parse(document)
|
|
let recovered = PureXML.read(document)
|
|
#expect(recovered.diagnostics.isEmpty, "unexpected diagnostics for: \(document) -> \(recovered.diagnostics)")
|
|
#expect(recovered.node == strict, "trees diverged for: \(document)")
|
|
}
|
|
}
|
|
|
|
@Test("The recovering editor tree projects to the same node as strict parse")
|
|
func test_readTreeAgrees() throws {
|
|
for document in corpus {
|
|
let strict = try PureXML.parse(document)
|
|
let (tree, diagnostics) = PureXML.readTree(document)
|
|
#expect(diagnostics.isEmpty, "unexpected diagnostics for: \(document)")
|
|
#expect(tree.node == strict, "readTree diverged for: \(document)")
|
|
}
|
|
}
|
|
}
|