[bugfix] XQuery 3.1 mandatory fixes from v2/xq4-core-functions (audit extract #3 subset)

exist-core/src/main/antlr/org/exist/xquery/parser/XQueryTree.g

@@ -1108,8 +1108,6 @@ throws XPathException
                 STAR
                 |
                 (
-                    // TODO: parameter types are collected, but not used!
-                    // Change SequenceType accordingly.
                     { List<SequenceType> paramTypes = new ArrayList<SequenceType>(5); }
                     (
                         { SequenceType paramType = new SequenceType(); }
@@ -1118,6 +1116,10 @@ throws XPathException
                     )*
                     { SequenceType returnType = new SequenceType(); }
                     "as" sequenceType [returnType]
+                    {
+                        type.setFunctionParamTypes(paramTypes.toArray(new SequenceType[0]));
+                        type.setFunctionReturnType(returnType);
+                    }
                 )
             )
         )
@@ -1128,14 +1130,15 @@ throws XPathException
                 STAR
                 |
                 (
-                    // TODO: parameter types are collected, but not used!
-                    // Change SequenceType accordingly.
                     { List<SequenceType> paramTypes = new ArrayList<SequenceType>(5); }
                     (
                         { SequenceType paramType = new SequenceType(); }
                         sequenceType [paramType]
                         { paramTypes.add(paramType); }
                     )*
+                    {
+                        type.setFunctionParamTypes(paramTypes.toArray(new SequenceType[0]));
+                    }
                 )
             )
         )
@@ -1146,14 +1149,15 @@ throws XPathException
                 STAR
                 |
                 (
-                    // TODO: parameter types are collected, but not used!
-                    // Change SequenceType accordingly.
                     { List<SequenceType> paramTypes = new ArrayList<SequenceType>(5); }
                     (
                         { SequenceType paramType = new SequenceType(); }
                         sequenceType [paramType]
                         { paramTypes.add(paramType); }
                     )*
+                    {
+                        type.setFunctionParamTypes(paramTypes.toArray(new SequenceType[0]));
+                    }
                 )
             )
         )

exist-core/src/main/java/org/exist/xquery/DynamicTypeCheck.java

@@ -149,6 +149,15 @@ private void check(Sequence result, Item item) throws XPathException {
             } else if (type == Type.ANY_URI && requiredType == Type.STRING) {
                     item = item.convertTo(Type.STRING);
                     type = Type.STRING;
+            //Binary type promotion (XQuery 4.0): xs:base64Binary ↔ xs:hexBinary
+            } else if ((type == Type.BASE64_BINARY && requiredType == Type.HEX_BINARY)
+                    || (type == Type.HEX_BINARY && requiredType == Type.BASE64_BINARY)) {
+                try {
+                    item = item.convertTo(requiredType);
+                } catch (final XPathException e) {
+                    throw new XPathException(expression, ErrorCodes.XPTY0004,
+                            "cannot convert " + Type.getTypeName(type) + " to " + Type.getTypeName(requiredType));
+                }
             } else {
                 if (!(Type.subTypeOf(type, requiredType))) {
                     throw new XPathException(expression, typeMismatchError,

exist-core/src/main/java/org/exist/xquery/ErrorCodes.java

@@ -90,6 +90,7 @@ public class ErrorCodes {
     public static final ErrorCode XQST0052 = new W3CErrorCode("XQST0052", "It is a static error if the type-name in a single-type or sequence-type for a cast or castable expression does not refer to a defined atomic type.");
     public static final ErrorCode XQST0053 = new W3CErrorCode("XQST0053", "(Not currently used.)");
     public static final ErrorCode XQST0054 = new W3CErrorCode("XQST0054", "It is a static error if a variable depends on itself.");
+    public static final ErrorCode XQDY0054 = new W3CErrorCode("XQDY0054", "It is a dynamic error if a variable depends on itself.");
     public static final ErrorCode XQST0055 = new W3CErrorCode("XQST0055", "It is a static error if a Prolog contains more than one copy-namespaces declaration.");
     public static final ErrorCode XQST0056 = new W3CErrorCode("XQST0056", "(Not currently used.)");
     public static final ErrorCode XQST0057 = new W3CErrorCode("XQST0057", "It is a static error if a schema import binds a namespace prefix but does not specify a target namespace other than a zero-length string.");

exist-core/src/main/java/org/exist/xquery/FunctionFactory.java

@@ -24,6 +24,7 @@
 import java.util.ArrayList;
 import java.util.Iterator;
 import java.util.List;
+import java.util.Set;
 
 import org.exist.Namespaces;
 import org.exist.dom.QName;
@@ -48,13 +49,37 @@ public class FunctionFactory {
     public static final String PROPERTY_DISABLE_DEPRECATED_FUNCTIONS = "xquery.disable-deprecated-functions";
     public static final boolean DISABLE_DEPRECATED_FUNCTIONS_BY_DEFAULT = false;
 
+    /**
+     * Reserved function names per XQuery 3.1/4.0 spec.
+     * These names must not be used as unprefixed function calls (XPST0003).
+     */
+    private static final Set<String> RESERVED_FUNCTION_NAMES = Set.of(
+        "array", "attribute", "comment", "document-node", "element",
+        "function", "if", "item", "map", "namespace-node", "node",
+        "processing-instruction", "schema-attribute", "schema-element",
+        "switch", "text", "typeswitch"
+    );
+
     public static Expression createFunction(XQueryContext context, XQueryAST ast, PathExpr parent, List<Expression> params) throws XPathException {
     	QName qname = null;
         try {
             qname = QName.parse(context, ast.getText(), context.getDefaultFunctionNamespace());
         } catch(final QName.IllegalQNameException xpe) {
             throw new XPathException(ast, ErrorCodes.XPST0081, "Invalid qname " +  ast.getText() + ". " + xpe.getMessage());
         }
+
+        // Check for reserved function names — unprefixed reserved names cannot be
+        // used as function calls (XPST0003). Prefixed names like fn:item() are not
+        // subject to the reserved name restriction (they just won't be found → XPST0017).
+        final String rawName = ast.getText();
+        if (rawName != null && !rawName.contains(":") && !rawName.contains("{")) {
+            final String local = qname.getLocalPart();
+            if (RESERVED_FUNCTION_NAMES.contains(local)) {
+                throw new XPathException(ast.getLine(), ast.getColumn(), ErrorCodes.XPST0003,
+                    "'" + local + "' is a reserved function name and cannot be used as a function call");
+            }
+        }
+
         return createFunction(context, qname, ast, parent, params);
     }
 
@@ -508,12 +533,13 @@ public static FunctionCall wrap(XQueryContext context, Function call) throws XPa
                 newSignature.setArgumentTypes(newParamArray);
 
 		final UserDefinedFunction func = new UserDefinedFunction(context, newSignature);
+		func.setPassContextToBody(true);
 		for (final QName varName: variables) {
 			func.addVariable(varName);
 		}
-		
+
 		call.setArguments(innerArgs);
-		
+
 		func.setFunctionBody(call);
 
 		final FunctionCall wrappedCall = new FunctionCall(context, func);

exist-core/src/main/java/org/exist/xquery/GeneralComparison.java

@@ -1082,7 +1082,17 @@ private AtomicValue convertForValueComparison(final AtomicValue value, final int
             }
 
             /*
-             *  d. Otherwise, a type error is raised [err:XPTY0004].
+             *  d. (XQuery 4.0) If each operand is an instance of one of the types
+             *     xs:hexBinary or xs:base64Binary, then both operands are cast to
+             *     type xs:base64Binary.
+             */
+            if ((thisType == Type.HEX_BINARY || thisType == Type.BASE64_BINARY)
+                    && (otherType == Type.HEX_BINARY || otherType == Type.BASE64_BINARY)) {
+                return value.convertTo(Type.BASE64_BINARY);
+            }
+
+            /*
+             *  e. Otherwise, a type error is raised [err:XPTY0004].
              */
             throw new XPathException(this, ErrorCodes.XPTY0004, "Incompatible primitive types");
         }

exist-core/src/main/java/org/exist/xquery/NamedFunctionReference.java

@@ -24,6 +24,8 @@
 import java.util.ArrayList;
 import java.util.List;
 
+import java.util.Set;
+
 import org.exist.dom.QName;
 import org.exist.xquery.parser.XQueryAST;
 import org.exist.xquery.util.ExpressionDumper;
@@ -52,7 +54,29 @@ public void analyze(AnalyzeContextInfo contextInfo) throws XPathException {
     	resolvedFunction.analyze(contextInfo);
 	}
 
+	/**
+	 * Reserved function names per XQuery 3.1/4.0 spec.
+	 * These names must not be used as unprefixed named function references (XPST0003).
+	 */
+	private static final Set<String> RESERVED_FUNCTION_NAMES = Set.of(
+		"array", "attribute", "comment", "document-node", "element",
+		"function", "if", "item", "map", "namespace-node", "node",
+		"processing-instruction", "schema-attribute", "schema-element",
+		"switch", "text", "typeswitch"
+	);
+
 	public static FunctionCall lookupFunction(Expression self, XQueryContext context, QName funcName, int arity) throws XPathException {
+		// Check for reserved function names — these cannot be used as named function references
+		final String localPart = funcName.getLocalPart();
+		final String nsURI = funcName.getNamespaceURI();
+		if (RESERVED_FUNCTION_NAMES.contains(localPart) &&
+				(nsURI == null || nsURI.isEmpty() ||
+				 Function.BUILTIN_FUNCTION_NS.equals(nsURI) ||
+				 context.getDefaultFunctionNamespace().equals(nsURI))) {
+			throw new XPathException(self, ErrorCodes.XPST0003,
+				"'" + localPart + "' is a reserved function name and cannot be used as a named function reference");
+		}
+
 		if (Function.BUILTIN_FUNCTION_NS.equals(funcName.getNamespaceURI())
 				&& "concat".equals(funcName.getLocalPart())
 				&& arity < 2) {

exist-core/src/main/java/org/exist/xquery/UserDefinedFunction.java

@@ -47,6 +47,7 @@ public class UserDefinedFunction extends Function implements Cloneable {
     private FunctionCall call;
     private boolean hasBeenReset = false;
     private List<ClosureVariable> closureVariables = null;
+    private boolean passContextToBody = false;
 
     public UserDefinedFunction(XQueryContext context, FunctionSignature signature) {
         super(context, signature);
@@ -60,6 +61,17 @@ public void setFunctionBody(Expression body) {
         this.body = body.simplify();
     }
 
+    /**
+     * Mark this UDF as a wrapper for an internal function (created by
+     * {@link FunctionFactory#wrap}). Wrapper functions pass the evaluation
+     * context through to their body so that context-dependent built-in
+     * functions (like fn:id, fn:idref, fn:string, etc.) can access the
+     * focus when called via function references.
+     */
+    public void setPassContextToBody(boolean passContext) {
+        this.passContextToBody = passContext;
+    }
+
     public void addVariable(final String varName) throws XPathException {
         try {
             final QName qname = QName.parse(context, varName, null);
@@ -155,7 +167,15 @@ public Sequence eval(Sequence contextSequence, Item contextItem) throws XPathExc
                             ", got " + currentArguments[j].getItemCount());
                 }
             }
-            result = body.eval(null, null);
+            // For wrapper functions (created by FunctionFactory.wrap for internal
+            // function references), pass the context through so context-dependent
+            // built-in functions can access the focus. For regular user-declared
+            // functions, the focus is absent per the XQuery spec.
+            if (passContextToBody) {
+                result = body.eval(contextSequence, contextItem);
+            } else {
+                result = body.eval(null, null);
+            }
             return result;
         } finally {
             // restore the local variable stack

exist-core/src/main/java/org/exist/xquery/VariableImpl.java

@@ -120,26 +120,26 @@ public void setSequenceType(SequenceType type) throws XPathException {
                 else {actualCardinality = Cardinality.EXACTLY_ONE;}
             	//Type.EMPTY is *not* a subtype of other types ; checking cardinality first
         		if (!getSequenceType().getCardinality().isSuperCardinalityOrEqualOf(actualCardinality))
-    				{throw new XPathException(getValue(), "XPTY0004: Invalid cardinality for variable $" + getQName() +
+    				{throw new XPathException(getValue(), ErrorCodes.XPTY0004, "Invalid cardinality for variable $" + getQName() +
     						". Expected " +
     						getSequenceType().getCardinality().getHumanDescription() +
     						", got " + actualCardinality.getHumanDescription());}
         		//TODO : ignore nodes right now ; they are returned as xs:untypedAtomicType
         		if (!Type.subTypeOf(getSequenceType().getPrimaryType(), Type.NODE)) {
             		if (!getValue().isEmpty() && !Type.subTypeOf(getValue().getItemType(), getSequenceType().getPrimaryType()))
-        				{throw new XPathException(getValue(), "XPTY0004: Invalid type for variable $" + getQName() +
+        				{throw new XPathException(getValue(), ErrorCodes.XPTY0004, "Invalid type for variable $" + getQName() +
         						". Expected " +
         						Type.getTypeName(getSequenceType().getPrimaryType()) +
         						", got " +Type.getTypeName(getValue().getItemType()));}
         		//Here is an attempt to process the nodes correctly
         		} else {
-        			//Same as above : we probably may factorize 
+        			//Same as above : we probably may factorize
             		if (!getValue().isEmpty() && !Type.subTypeOf(getValue().getItemType(), getSequenceType().getPrimaryType()))
-        				{throw new XPathException(getValue(), "XPTY0004: Invalid type for variable $" + getQName() +
+        				{throw new XPathException(getValue(), ErrorCodes.XPTY0004, "Invalid type for variable $" + getQName() +
         						". Expected " +
         						Type.getTypeName(getSequenceType().getPrimaryType()) +
         						", got " +Type.getTypeName(getValue().getItemType()));}
-        			
+
         		}
             }
 

exist-core/src/main/java/org/exist/xquery/VariableReference.java

@@ -64,7 +64,7 @@ public void analyze(final AnalyzeContextInfo contextInfo) throws XPathException
                     "Variable '$" + qname + "' is not declared.");
         }
         if (!var.isInitialized()) {
-            throw new XPathException(this, ErrorCodes.XQST0054,
+            throw new XPathException(this, ErrorCodes.XQDY0054,
                     "variable declaration of '$" + qname + "' cannot " +
                             "be executed because of a circularity.");
         }

exist-core/src/main/java/org/exist/xquery/XQueryContext.java

@@ -1346,6 +1346,31 @@ public DocumentSet getStaticDocs() {
         return textResourceSupplier.apply(getBroker(), getBroker().getCurrentTransaction(), uri, charset);
     }
 
+    /**
+     * Gets a text resource from the "Available text resources" of the
+     * dynamic context, matching by URI only. This is used when no encoding
+     * is specified, allowing the resource to be found regardless of what
+     * charset it was registered with.
+     *
+     * @param uri the URI of the resource to retrieve
+     * @return a reader to read the resource content from, or null if not found
+     * @throws XPathException in case of a dynamic error
+     */
+    public @Nullable Reader getDynamicallyAvailableTextResourceByUri(final String uri)
+            throws XPathException {
+        if (dynamicTextResources == null) {
+            return null;
+        }
+
+        for (final Map.Entry<Tuple2<String, Charset>, QuadFunctionE<DBBroker, Txn, String, Charset, Reader, XPathException>> entry : dynamicTextResources.entrySet()) {
+            if (entry.getKey()._1.equals(uri)) {
+                final Charset registeredCharset = entry.getKey()._2;
+                return entry.getValue().apply(getBroker(), getBroker().getCurrentTransaction(), uri, registeredCharset);
+            }
+        }
+        return null;
+    }
+
     /**
      * Gets a collection from the "Available collections" of the
      * dynamic context.

exist-core/src/main/java/org/exist/xquery/functions/fn/FnModule.java

@@ -78,6 +78,7 @@ public class FnModule extends AbstractInternalModule {
         new FunctionDef(FunDocumentURI.FS_DOCUMENT_URI_1, FunDocumentURI.class),
         new FunctionDef(FunElementWithId.FS_ELEMENT_WITH_ID_SIGNATURES[0], FunElementWithId.class),
         new FunctionDef(FunElementWithId.FS_ELEMENT_WITH_ID_SIGNATURES[1], FunElementWithId.class),
+        new FunctionDef(FunElementWithId.FS_ELEMENT_WITH_ID_SIGNATURES[2], FunElementWithId.class),
         new FunctionDef(FunEmpty.signature, FunEmpty.class),
         new FunctionDef(FunEncodeForURI.signature, FunEncodeForURI.class),
         new FunctionDef(FunEndsWith.signatures[0], FunEndsWith.class),

exist-core/src/main/java/org/exist/xquery/functions/fn/FunAnalyzeString.java

@@ -125,9 +125,23 @@
         }
     }
 
-    private void analyzeString(final MemTreeBuilder builder, final String input, String pattern, final String flags) throws XPathException {
+    private void analyzeString(final MemTreeBuilder builder, final String input, final String pattern, final String flags) throws XPathException {
         final Configuration config = context.getBroker().getBrokerPool().getSaxonConfiguration();
 
+        // XPath 4.0 lookaround syntax is not yet implemented in eXist's XQuery 3.1 runtime.
+        // When XQuery 4.0 lands (v2/xq4-core-functions), replace this guard with the
+        // translateXPath4Lookaround() dispatch path.
+        if (org.exist.xquery.regex.RegexUtil.hasXPath4Lookaround(pattern)) {
+            throw new XPathException(this, ErrorCodes.XPST0017,
+                    "XPath 4.0 lookaround syntax in regex patterns (e.g. (*positive_lookahead:...)) "
+                            + "is not yet implemented in this XQuery 3.1 build. Rewrite the regex without lookaround.");
+        }
+
+        // Pre-validate: reject constructs not valid in XPath 3.1 regex
+        if (!org.exist.xquery.regex.RegexUtil.hasLiteral(flags)) {
+            org.exist.xquery.regex.RegexUtil.validateXPathRegex(this, pattern, false);
+        }
+
         final List<String> warnings = new ArrayList<>(1);
 
         try {

exist-core/src/main/java/org/exist/xquery/functions/fn/FunContainsToken.java

@@ -44,7 +44,7 @@ public class FunContainsToken extends BasicFunction {
 
     private final static FunctionParameterSequenceType FS_INPUT = optManyParam("input", Type.STRING, "The input string");
     private final static FunctionParameterSequenceType FS_TOKEN = param("token", Type.STRING, "The token to be searched for");
-    private final static FunctionParameterSequenceType FS_COLLATION = param("pattern", Type.STRING, "Collation to use");
+    private final static FunctionParameterSequenceType FS_COLLATION = optParam("collation", Type.STRING, "Collation to use; an empty sequence selects the default collation");
 
     public final static FunctionSignature[] FS_CONTAINS_TOKEN = functionSignatures(
             FS_CONTAINS_TOKEN_NAME,

exist-core/src/main/java/org/exist/xquery/functions/fn/FunElementWithId.java

@@ -45,13 +45,15 @@ public class FunElementWithId extends BasicFunction {
                     "If none is matching or $idrefs is the empty sequence, returns the empty sequence.";
     private static final FunctionReturnSequenceType FN_RETURN = returnsOptMany(Type.STRING, "the elements with IDs matching IDREFs from $idref-sequence");
     private static final FunctionParameterSequenceType PARAM_ID_REFS_STRING = optManyParam("idrefs", Type.STRING, "The IDREF sequence");
+    private static final FunctionParameterSequenceType PARAM_NODE = param("node", Type.NODE, "A node in the document to search");
     public static final FunctionSignature[] FS_ELEMENT_WITH_ID_SIGNATURES = functionSignatures(
             FN_NAME,
             FN_DESCRIPTION,
             FN_RETURN,
             arities(
                     arity(),
-                    arity(PARAM_ID_REFS_STRING)
+                    arity(PARAM_ID_REFS_STRING),
+                    arity(PARAM_ID_REFS_STRING, PARAM_NODE)
             )
     );