Mihaela Mihaljevic
380160cc2b
The two monolithic XSD content/identity rules only satisfied the OpenAPIKit validation idiom at the wrong granularity: no per-constraint removability, no isolable failing/near-miss tests. Split them into 32 named rules, each finding tagged at its origin and cached per XSDContext so the walk still runs once per document (the #101 decision against duplicating XSD's recursive type resolution still holds). Wired the streaming path's ID/IDREF and identity checks through the same framework instead of calling the validators imperatively, and added the unwrapAndLookup combinator the rule already named as required. Also includes the mandated repo-wide swiftformat/swiftlint pass.
401 lines
22 KiB
Swift
401 lines
22 KiB
Swift
public extension PureXML.Schema {
|
|
/// An error compiling or applying a schema.
|
|
enum SchemaError: Swift.Error, Equatable, Sendable, CustomStringConvertible {
|
|
case notASchema
|
|
/// A type inside `xs:redefine` does not redefine itself: its restriction or
|
|
/// extension base must name the type it redefines.
|
|
case redefineIncompatible(type: String)
|
|
/// An `xs:all` group violates its XSD 1.0 constraints: every member must be
|
|
/// an element with `maxOccurs` at most 1, and the group itself may occur at
|
|
/// most once.
|
|
case invalidAllGroup(reason: String)
|
|
/// A RELAX NG schema document does not match the RELAX NG grammar or
|
|
/// its restrictions (sections 3, 4.16-4.18).
|
|
case invalidRelaxNG(reason: String)
|
|
/// The compiled schema fails one or more consistency rules (a `final`
|
|
/// violation, an unfaithful restriction); every finding is carried, so a
|
|
/// schema with several problems reports them all at once.
|
|
case inconsistent([PureXML.Validation.ValidationError])
|
|
|
|
public var description: String {
|
|
switch self {
|
|
case .notASchema:
|
|
"the document is not an xs:schema"
|
|
case let .redefineIncompatible(type):
|
|
"redefined type '\(type)' must derive from itself"
|
|
case let .invalidAllGroup(reason):
|
|
"invalid xs:all group: \(reason)"
|
|
case let .invalidRelaxNG(reason):
|
|
"invalid RELAX NG schema: \(reason)"
|
|
case let .inconsistent(findings):
|
|
findings.map(\.description).joined(separator: "; ")
|
|
}
|
|
}
|
|
|
|
/// Every compile-time finding when the schema is inconsistent.
|
|
public var inconsistentFindings: [PureXML.Validation.ValidationError]? {
|
|
if case let .inconsistent(findings) = self {
|
|
return findings
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
|
|
/// A compiled XSD schema: its global element declarations and named-type
|
|
/// table, parsed from a schema document, used to validate instance documents.
|
|
struct Document: Sendable {
|
|
private let elements: [String: ElementType]
|
|
private let types: [String: ElementType]
|
|
private let constraints: [String: [IdentityConstraint]]
|
|
private let identityFieldTypes: [String: SimpleType]
|
|
private let identityFieldConstraints: [String: ValueConstraint]
|
|
private let identityNonSimpleFields: Set<String>
|
|
private let nillableElements: Set<String>
|
|
private let elementConstraints: [String: ValueConstraint]
|
|
private let abstractTypes: Set<String>
|
|
private let abstractElements: Set<String>
|
|
private let typeBlock: [String: Set<DerivationMethod>]
|
|
private let elementBlock: [String: Set<DerivationMethod>]
|
|
private let typeDerivation: [String: TypeDerivation]
|
|
private let unionMembers: [String: [String]]
|
|
private let targetNamespace: String?
|
|
private let globalAttributes: [String: AttributeUse]
|
|
|
|
/// Combines already-compiled declaration tables, for merging the schemas
|
|
/// an instance references through `xsi:schemaLocation`. No consistency
|
|
/// check runs: each component schema was validated when first compiled.
|
|
private init(
|
|
elements: [String: ElementType],
|
|
types: [String: ElementType],
|
|
constraints: [String: [IdentityConstraint]],
|
|
identityFieldTypes: [String: SimpleType] = [:],
|
|
identityFieldConstraints: [String: ValueConstraint] = [:],
|
|
identityNonSimpleFields: Set<String> = [],
|
|
nillableElements: Set<String>,
|
|
elementConstraints: [String: ValueConstraint],
|
|
abstractTypes: Set<String>,
|
|
abstractElements: Set<String>,
|
|
typeBlock: [String: Set<DerivationMethod>],
|
|
elementBlock: [String: Set<DerivationMethod>],
|
|
typeDerivation: [String: TypeDerivation],
|
|
unionMembers: [String: [String]],
|
|
targetNamespace: String?,
|
|
globalAttributes: [String: AttributeUse],
|
|
) {
|
|
self.elements = elements
|
|
self.types = types
|
|
self.constraints = constraints
|
|
self.identityFieldTypes = identityFieldTypes
|
|
self.identityFieldConstraints = identityFieldConstraints
|
|
self.identityNonSimpleFields = identityNonSimpleFields
|
|
self.nillableElements = nillableElements
|
|
self.elementConstraints = elementConstraints
|
|
self.abstractTypes = abstractTypes
|
|
self.abstractElements = abstractElements
|
|
self.typeBlock = typeBlock
|
|
self.elementBlock = elementBlock
|
|
self.typeDerivation = typeDerivation
|
|
self.unionMembers = unionMembers
|
|
self.targetNamespace = targetNamespace
|
|
self.globalAttributes = globalAttributes
|
|
}
|
|
|
|
/// This schema with `other`'s global declarations merged in; this schema's
|
|
/// own declarations win on any key conflict. Used to fold in the schemas an
|
|
/// instance points at, so a strict/lax wildcard can resolve elements
|
|
/// declared in another document.
|
|
public func merged(with other: Document) -> Document {
|
|
merging(other)
|
|
}
|
|
|
|
private func merging(_ other: Document) -> Document {
|
|
func mineFirst<V>(_ mine: [String: V], _ theirs: [String: V]) -> [String: V] {
|
|
mine.merging(theirs) { current, _ in current }
|
|
}
|
|
return Document(
|
|
elements: mineFirst(elements, other.elements),
|
|
types: mineFirst(types, other.types),
|
|
constraints: mineFirst(constraints, other.constraints),
|
|
identityFieldTypes: mineFirst(identityFieldTypes, other.identityFieldTypes),
|
|
identityFieldConstraints: mineFirst(identityFieldConstraints, other.identityFieldConstraints),
|
|
identityNonSimpleFields: identityNonSimpleFields.union(other.identityNonSimpleFields),
|
|
nillableElements: nillableElements.union(other.nillableElements),
|
|
elementConstraints: mineFirst(elementConstraints, other.elementConstraints),
|
|
abstractTypes: abstractTypes.union(other.abstractTypes),
|
|
abstractElements: abstractElements.union(other.abstractElements),
|
|
typeBlock: mineFirst(typeBlock, other.typeBlock),
|
|
elementBlock: mineFirst(elementBlock, other.elementBlock),
|
|
typeDerivation: mineFirst(typeDerivation, other.typeDerivation),
|
|
unionMembers: mineFirst(unionMembers, other.unionMembers),
|
|
targetNamespace: targetNamespace,
|
|
globalAttributes: mineFirst(globalAttributes, other.globalAttributes),
|
|
)
|
|
}
|
|
|
|
/// Compiles a schema document. `schemaLoader` resolves the
|
|
/// `schemaLocation` of `xs:include`, `xs:import`, and `xs:redefine` to
|
|
/// schema source; it returns nil (the default) when external schemas are
|
|
/// not available, which keeps compilation from reaching the filesystem or
|
|
/// network by default.
|
|
private init(compiled: XSDCompiled) throws {
|
|
elements = compiled.elements
|
|
types = compiled.types
|
|
constraints = compiled.constraints
|
|
identityFieldTypes = compiled.identityFieldTypes
|
|
identityNonSimpleFields = compiled.identityNonSimpleFields
|
|
identityFieldConstraints = compiled.identityFieldConstraints
|
|
nillableElements = compiled.nillableElements
|
|
elementConstraints = compiled.elementConstraints
|
|
abstractTypes = compiled.nsAbstractTypes
|
|
abstractElements = compiled.abstractElements
|
|
typeBlock = compiled.nsTypeBlock
|
|
elementBlock = compiled.nsElementBlock
|
|
typeDerivation = compiled.nsTypeDerivation
|
|
unionMembers = compiled.nsUnionMembers
|
|
targetNamespace = compiled.targetNamespace
|
|
globalAttributes = compiled.globalAttributes
|
|
// Schema consistency through the validation framework: every named
|
|
// type is checked by the composable rules (final respected, Particle
|
|
// Valid (Restriction)), and ALL findings are reported together, joined
|
|
// with the schema-validity findings gathered while parsing (malformed
|
|
// facet definitions and the like).
|
|
let consistencyFindings = PureXML.Validation.XSDSchema.consistencyErrors(
|
|
types: compiled.types,
|
|
typeDerivation: compiled.typeDerivation,
|
|
typeFinal: compiled.typeFinal,
|
|
)
|
|
let allFindings = compiled.schemaErrors + consistencyFindings
|
|
if !allFindings.isEmpty {
|
|
throw SchemaError.inconsistent(allFindings)
|
|
}
|
|
}
|
|
|
|
/// Validates an instance document against the schema, returning one located
|
|
/// ``PureXML/Validation/ValidationError`` per violation. Reports an error
|
|
/// when the root element has no global declaration.
|
|
public func validate(_ xml: String) throws -> [PureXML.Validation.ValidationError] {
|
|
try validate(PureXML.parse(xml))
|
|
}
|
|
|
|
/// Validates `xml`, additionally honoring the instance's
|
|
/// `xsi:schemaLocation` and `xsi:noNamespaceSchemaLocation` hints: each
|
|
/// referenced schema document is loaded through `schemaLoader`, compiled,
|
|
/// and its global declarations merged in (this schema's own declarations
|
|
/// win), so a strict or lax wildcard can resolve elements declared in
|
|
/// another document. With no hints this is `validate(_:)`.
|
|
public func validate(_ xml: String, schemaLoader: @escaping (String) -> String?) throws -> [PureXML.Validation.ValidationError] {
|
|
let node = try PureXML.parse(xml)
|
|
return merged(with: node, schemaLoader: schemaLoader).validate(node)
|
|
}
|
|
|
|
/// Merges global declarations from every schema document the instance
|
|
/// references through `xsi:schemaLocation` or
|
|
/// `xsi:noNamespaceSchemaLocation`.
|
|
private func merged(with node: PureXML.Model.Node, schemaLoader: @escaping (String) -> String?) -> Document {
|
|
var combined = self
|
|
for location in Self.hintedSchemaLocations(in: node) {
|
|
guard let source = schemaLoader(location),
|
|
let other = try? Document(source, schemaLoader: schemaLoader)
|
|
else { continue }
|
|
combined = combined.merging(other)
|
|
}
|
|
return combined
|
|
}
|
|
|
|
/// Validates `xml` against the schema while it is pulled event by event (the
|
|
/// libxml2 `xmlTextReader` model) rather than over a built tree. Memory is
|
|
/// bounded to the open-element stack. Content models, attributes, and simple
|
|
/// content are checked incrementally; document-scoped identity constraints
|
|
/// are checked once at the end from a parsed tree (the instance is parsed
|
|
/// again only when constraints are present).
|
|
public func validate(streaming xml: String, limits: PureXML.Parsing.Limits = .default) throws -> [PureXML.Validation.ValidationError] {
|
|
let identityNode = constraints.isEmpty ? nil : try PureXML.parse(xml)
|
|
return try validateStreaming(xml, limits: limits, identityNode: identityNode)
|
|
}
|
|
|
|
/// Validates `xml` while it is pulled event by event, additionally honoring
|
|
/// the instance's `xsi:schemaLocation` and
|
|
/// `xsi:noNamespaceSchemaLocation` hints through `schemaLoader`. See
|
|
/// ``validate(_:schemaLoader:)-(String)``.
|
|
public func validate(
|
|
streaming xml: String,
|
|
schemaLoader: @escaping (String) -> String?,
|
|
limits: PureXML.Parsing.Limits = .default,
|
|
) throws -> [PureXML.Validation.ValidationError] {
|
|
let node = try PureXML.parse(xml)
|
|
return try merged(with: node, schemaLoader: schemaLoader).validateStreaming(xml, limits: limits, identityNode: node)
|
|
}
|
|
|
|
private func validateStreaming(
|
|
_ xml: String,
|
|
limits: PureXML.Parsing.Limits,
|
|
identityNode: PureXML.Model.Node?,
|
|
) throws -> [PureXML.Validation.ValidationError] {
|
|
let validator = PureXML.Schema.ComplexValidator(
|
|
types: types,
|
|
globalAttributes: globalAttributes,
|
|
nillableElements: nillableElements,
|
|
elementConstraints: elementConstraints,
|
|
abstractTypes: abstractTypes,
|
|
typeBlock: typeBlock,
|
|
elementBlock: elementBlock,
|
|
typeDerivation: typeDerivation,
|
|
unionMembers: unionMembers,
|
|
)
|
|
var driver = PureXML.Validation.StreamingXSDValidator(
|
|
validator: validator,
|
|
rootElements: elements,
|
|
abstractElements: abstractElements,
|
|
targetNamespace: targetNamespace,
|
|
)
|
|
var reader = PureXML.Parsing.EventReader(xml, limits: limits)
|
|
while let event = try reader.next() {
|
|
driver.consume(event)
|
|
}
|
|
var errors = driver.finish()
|
|
// Document-scoped ID/IDREF and identity-constraint checking run once the
|
|
// whole document has streamed through; both go through the same named
|
|
// ``PureXML/Validation`` rules the tree path uses (``idConsistency``,
|
|
// `XSD.identityConstraintRules`), not a bare imperative call, so the
|
|
// streaming path stays inside the framework like every other validator.
|
|
errors += PureXML.Validation.Validator<Void>.blank
|
|
.validating(PureXML.Schema.ComplexValidator.idConsistency)
|
|
.errors(for: validator, in: ())
|
|
if !constraints.isEmpty, let identityNode {
|
|
errors += identityErrors(in: identityNode)
|
|
}
|
|
return errors
|
|
}
|
|
|
|
/// Identity-constraint findings for an already-parsed instance tree, run
|
|
/// through the same decomposed ``PureXML/Validation/XSD`` identity rules
|
|
/// the tree path uses.
|
|
private func identityErrors(in node: PureXML.Model.Node) -> [PureXML.Validation.ValidationError] {
|
|
guard case let .document(children) = node, let root = children.compactMap(\.element).first else {
|
|
return []
|
|
}
|
|
let context = PureXML.Validation.XSDContext(
|
|
types: types,
|
|
constraints: constraints,
|
|
rootDeclaration: nil,
|
|
identityFieldTypes: identityFieldTypes,
|
|
identityFieldConstraints: identityFieldConstraints,
|
|
identityNonSimpleFields: identityNonSimpleFields,
|
|
)
|
|
return PureXML.Validation.Validator<PureXML.Validation.XSDContext>.defaults(
|
|
reference: PureXML.Validation.XSD.identityConstraintRules.map { PureXML.Validation.AnyValidation($0) },
|
|
)
|
|
.errors(for: .element(root), in: context)
|
|
}
|
|
|
|
/// Validates an already-parsed node tree, so a caller can validate the same
|
|
/// (possibly recovered) tree it is editing. See ``validate(_:)-(String)``.
|
|
public func validate(_ node: PureXML.Model.Node) -> [PureXML.Validation.ValidationError] {
|
|
guard case let .document(children) = node, let root = children.compactMap(\.element).first else {
|
|
return [.init(reason: "the document has no root element", at: [])]
|
|
}
|
|
guard let declaration = globalElementDeclaration(for: root.name) ?? rootTypeOverride(for: root) else {
|
|
return [.init(reason: "no element declaration for '\(root.name.localName)'", at: [.element(root.name.description)])]
|
|
}
|
|
// A schema with a target namespace declares its global elements in that
|
|
// namespace; the root must be in it unless a merged schema declares it
|
|
// in another namespace, or xsi:type names the validating type.
|
|
if rootTypeOverride(for: root) == nil, rootNamespaceMismatch(root.name) {
|
|
let reason = "root element '\(root.name.localName)' is not in the schema target namespace '\(targetNamespace ?? "")'"
|
|
return [.init(reason: reason, at: [.element(root.name.description)])]
|
|
}
|
|
// An abstract element may not appear in an instance directly; only a
|
|
// concrete substitution-group member may stand in its place.
|
|
if abstractElements.contains(root.name.localName) {
|
|
return [.init(reason: "abstract element '\(root.name.localName)' must not appear in an instance", at: [.element(root.name.description)])]
|
|
}
|
|
let context = PureXML.Validation.XSDContext(
|
|
types: types,
|
|
constraints: constraints,
|
|
rootDeclaration: declaration,
|
|
nillableElements: nillableElements,
|
|
elementConstraints: elementConstraints,
|
|
abstractTypes: abstractTypes,
|
|
typeBlock: typeBlock,
|
|
elementBlock: elementBlock,
|
|
typeDerivation: typeDerivation,
|
|
unionMembers: unionMembers,
|
|
globalAttributes: globalAttributes,
|
|
identityFieldTypes: identityFieldTypes,
|
|
identityFieldConstraints: identityFieldConstraints,
|
|
identityNonSimpleFields: identityNonSimpleFields,
|
|
)
|
|
return PureXML.Validation.XSD.validator().errors(for: .element(root), in: context)
|
|
}
|
|
|
|
/// The global element declaration for a qualified instance name.
|
|
private func globalElementDeclaration(for name: PureXML.Model.QualifiedName) -> ElementType? {
|
|
if let declaration = types[PureXML.Schema.XSDParser.elementDeclarationKey(name)] {
|
|
return declaration
|
|
}
|
|
if name.namespaceURI == nil || name.namespaceURI?.isEmpty == true {
|
|
return elements[name.localName]
|
|
}
|
|
return nil
|
|
}
|
|
|
|
private func rootNamespaceMismatch(_ name: PureXML.Model.QualifiedName) -> Bool {
|
|
guard let target = targetNamespace, !target.isEmpty, name.namespaceURI != target else { return false }
|
|
return types[PureXML.Schema.XSDParser.elementDeclarationKey(name)] == nil
|
|
}
|
|
|
|
/// When the root has no global element declaration, an `xsi:type` attribute
|
|
/// can still name the type to validate against (Sun target-NS tests).
|
|
private func rootTypeOverride(for root: PureXML.Model.Element) -> ElementType? {
|
|
guard PureXML.Schema.ComplexValidator.xsiTypeAttributeValue(root) != nil else { return nil }
|
|
let bindings = PureXML.Schema.ComplexValidator.namespaceBindings(for: root)
|
|
guard let reference = PureXML.Schema.ComplexValidator.xsiTypeReference(root, namespaceBindings: bindings),
|
|
let type = PureXML.Schema.ComplexValidator.resolveNamedType(reference, in: types)
|
|
else { return nil }
|
|
return type
|
|
}
|
|
|
|
/// What the schema allows at the element a coding `path` addresses in
|
|
/// `tree`: the next child elements (the content-model follow-set), whether
|
|
/// the content may end there, and the declared attributes with their
|
|
/// required/present status. For editor completions and "what's missing".
|
|
/// Returns nil when the path does not address a declared element.
|
|
public func completions(at path: [PureXML.Validation.PathKey], in tree: PureXML.Model.TreeNode) -> Completions? {
|
|
guard let node = tree.node(at: path), case let .element(element) = node.node,
|
|
let type = CompletionEngine.elementType(at: path, elements: elements, types: types)
|
|
else {
|
|
return nil
|
|
}
|
|
return CompletionEngine.completions(for: element, type: type, types: types)
|
|
}
|
|
|
|
/// The quick-fixes the schema offers at the element a coding `path`
|
|
/// addresses in `tree`: add a required-but-absent attribute, or insert a
|
|
/// still-expected required child. Derived from the structured
|
|
/// ``completions(at:in:)``, with precise placement from the node's content
|
|
/// span, so the edits are exact, not guessed from a message.
|
|
public func quickFixes(at path: [PureXML.Validation.PathKey], in tree: PureXML.Model.TreeNode) -> [PureXML.QuickFix] {
|
|
guard let completions = completions(at: path, in: tree), let element = tree.node(at: path) else { return [] }
|
|
return PureXML.QuickFixEngine.fixes(from: completions, element: element)
|
|
}
|
|
}
|
|
}
|
|
|
|
public extension PureXML.Schema.Document {
|
|
/// Compiles a schema document. `schemaLoader` resolves the `schemaLocation` of
|
|
/// `xs:include`, `xs:import`, and `xs:redefine` to schema source; it returns
|
|
/// nil (the default) when external schemas are not available, which keeps
|
|
/// compilation from reaching the filesystem or network by default.
|
|
init(_ xsd: String, schemaLoader: @escaping (String) -> String? = { _ in nil }) throws {
|
|
try self.init(compiled: PureXML.Schema.XSDParser.parse(xsd, loader: schemaLoader))
|
|
}
|
|
|
|
/// Compiles several schema documents that jointly form one schema as a single
|
|
/// union (#161), so cross-document substitution-group membership is resolved.
|
|
/// Use this rather than compiling each document and merging the results, which
|
|
/// freezes each document's content models before the others are known.
|
|
init(composing sources: [String], schemaLoader: @escaping (String) -> String? = { _ in nil }) throws {
|
|
try self.init(compiled: PureXML.Schema.XSDParser.parse(union: sources, loader: schemaLoader))
|
|
}
|
|
}
|