001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *      https://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017
018package org.apache.commons.xml.secure;
019
020import java.lang.invoke.MethodHandle;
021import java.util.Objects;
022
023import javax.xml.XMLConstants;
024import javax.xml.xpath.XPath;
025import javax.xml.xpath.XPathFactory;
026import javax.xml.xpath.XPathFactoryConfigurationException;
027import javax.xml.xpath.XPathFunctionResolver;
028import javax.xml.xpath.XPathVariableResolver;
029
030/**
031 * Creates new, secure {@link XPathFactory} instances.
032 * <p>
033 * Beyond the three universal guarantees on {@link org.apache.commons.xml.secure}, URI-fetching XPath 3.1+ functions ({@code doc()}, {@code collection()},
034 * {@code unparsed-text()}) are not resolved.
035 * </p>
036 * <p>
037 * The guarantees also cover the document parse behind {@code XPath.evaluate(String, InputSource)} and {@code XPathExpression.evaluate(InputSource)}: the input
038 * document is built through a secure, namespace-aware {@link javax.xml.parsers.DocumentBuilder} instead of the engine's internal parser.
039 * </p>
040 * <p>
041 * This class is not itself an {@link XPathFactory}, so it inherits none of the static JAXP factory methods. A caller therefore cannot obtain an unsecured
042 * factory through this class by calling a method such as {@code newDefaultInstance()}. The secure factories are instances of a nested, non-public wrapper
043 * class.
044 * </p>
045 *
046 * @see org.apache.commons.xml.secure
047 */
048public final class SecureXPathFactory {
049
050    /**
051     * {@link XPathFactory} wrapper that returns a {@link SecureXPath} from {@link #newXPath()}.
052     * <p>
053     * Required because {@link javax.xml.XMLConstants#FEATURE_SECURE_PROCESSING} on the factory governs only the XPath engine: the stock JDK and Apache Xalan
054     * implement the {@link org.xml.sax.InputSource}-taking {@code evaluate} entry points by provisioning an internal document parser the feature does not
055     * reach. The wrapper performs that document build itself through a secure parser instead; see {@link SecureXPath}.
056     * </p>
057     */
058    private static final class Wrapper extends XPathFactory {
059
060        private final XPathFactory delegate;
061
062        /**
063         * Constructs a new instance.
064         *
065         * @param delegate The delegate to wrap; must not be {@code null}.
066         * @throws NullPointerException Thrown if {@code delegate} is {@code null}.
067         */
068        private Wrapper(final XPathFactory delegate) {
069            this.delegate = Objects.requireNonNull(delegate, "delegate");
070        }
071
072        @Override
073        public boolean getFeature(final String name) throws XPathFactoryConfigurationException {
074            return delegate.getFeature(name);
075        }
076
077        /**
078         * Gets a property of the delegate through the Java 18 {@code XPathFactory.getProperty(String)} method.
079         * <p>
080         * Not marked {@code @Override}: this library compiles against the Java 8 API, where {@link XPathFactory} declares no such method, so the annotation
081         * would not compile. At run time on Java 18 or later, it overrides the inherited method, which would otherwise answer for the wrapper and hide the
082         * delegate's own limits ({@code jdk.xml.xpath*}) behind an {@code UnsupportedOperationException}.
083         * </p>
084         *
085         * @param name The property name.
086         * @return the delegate's value for the property.
087         */
088        public String getProperty(final String name) {
089            if (MH_getProperty == null) {
090                throw new UnsupportedOperationException("XPathFactory.getProperty(String) requires Java 18 or later");
091            }
092            return MethodHandleFactory.invokeExact(() -> (String) MH_getProperty.invokeExact(delegate, name), RuntimeException.class);
093        }
094
095        @Override
096        public boolean isObjectModelSupported(final String objectModel) {
097            return delegate.isObjectModelSupported(objectModel);
098        }
099
100        @Override
101        public XPath newXPath() {
102            // newXPath() should never return null for a specification-compliant factory.
103            final XPath xpath = delegate.newXPath();
104            return xpath == null ? null : new SecureXPath(xpath, overrideDefaultParser());
105        }
106
107        /**
108         * Tests whether parsers should be instantiated via {@code newInstance()} instead of {@code newDefaultInstance()}.
109         * <p>
110         * The JDK implementation of {@link XPathFactory} uses the JDK parsers while {@value SecureSAXParserFactory#OVERRIDE_DEFAULT_PARSER} is unset or
111         * {@code false}.
112         * </p>
113         *
114         * @return {@code true} if parsers should be created via {@code newInstance()}.
115         */
116        private boolean overrideDefaultParser() {
117            try {
118                return delegate.getFeature(SecureSAXParserFactory.OVERRIDE_DEFAULT_PARSER);
119            } catch (final XPathFactoryConfigurationException e) {
120                return true;
121            }
122        }
123
124        @Override
125        public void setFeature(final String name, final boolean value) throws XPathFactoryConfigurationException {
126            delegate.setFeature(name, value);
127        }
128
129        /**
130         * Sets a property on the delegate through the Java 18 {@code XPathFactory.setProperty(String, String)} method.
131         *
132         * <p>
133         * See {@link #getProperty(String)} for why it carries no {@code @Override}. The {@code jdk.xml.xpath*} limits reached this way are processing limits
134         * like any other: an operator may tighten them, and loosening one is reconfiguration.
135         * </p>
136         *
137         * @param name  The property name.
138         * @param value The value to set.
139         */
140        public void setProperty(final String name, final String value) {
141            if (MH_setProperty == null) {
142                throw new UnsupportedOperationException("XPathFactory.setProperty(String, String) requires Java 18 or later");
143            }
144            MethodHandleFactory.invokeExact(() -> {
145                MH_setProperty.invokeExact(delegate, name, value);
146                return null;
147            }, RuntimeException.class);
148        }
149
150        @Override
151        public void setXPathFunctionResolver(final XPathFunctionResolver resolver) {
152            delegate.setXPathFunctionResolver(resolver);
153        }
154
155        @Override
156        public void setXPathVariableResolver(final XPathVariableResolver resolver) {
157            delegate.setXPathVariableResolver(resolver);
158        }
159    }
160
161    /**
162     * Class name of the JDK's built-in default implementation, the Java 8 fallback for {@link #newDefaultInstance()}.
163     */
164    private static final String JDK_XPATH_FACTORY = "com.sun.org.apache.xpath.internal.jaxp.XPathFactoryImpl";
165
166    private static final MethodHandle MH_newDefaultInstance = MethodHandleFactory.findStatic(XPathFactory.class, "newDefaultInstance");
167
168    /**
169     * {@code XPathFactory.getProperty(String)}, added in Java 18; {@code null} on earlier releases, where the method does not exist.
170     */
171    private static final MethodHandle MH_getProperty = MethodHandleFactory.findVirtual(XPathFactory.class, "getProperty", String.class, String.class);
172
173    /**
174     * {@code XPathFactory.setProperty(String, String)}, added in Java 18; {@code null} on earlier releases, where the method does not exist.
175     */
176    private static final MethodHandle MH_setProperty =
177            MethodHandleFactory.findVirtual(XPathFactory.class, "setProperty", void.class, String.class, String.class);
178
179    /**
180     * Returns a new, secure {@link XPathFactory} of the system-default implementation, supporting the default XPath object model.
181     * <p>
182     * Obtained from {@code XPathFactory.newDefaultInstance()} where the platform provides it (Java 9 or later), and by instantiating the JDK's built-in
183     * implementation directly on Java 8.
184     * </p>
185     *
186     * @return A secure factory.
187     * @throws IllegalStateException Thrown if a required secure setting cannot be applied to the underlying implementation.
188     * @throws RuntimeException      Thrown if the running platform provides neither {@code newDefaultInstance()} nor the JDK's built-in implementation (for
189     *                               example Android).
190     */
191    public static XPathFactory newDefaultInstance() {
192        if (MH_newDefaultInstance != null) {
193            return secure(MethodHandleFactory.invokeExact(() -> (XPathFactory) MH_newDefaultInstance.invokeExact(), RuntimeException.class));
194        }
195        try {
196            // Java 8: the method does not exist; instantiate the JDK's built-in default by its class name instead.
197            return newInstance(XPathFactory.DEFAULT_OBJECT_MODEL_URI, JDK_XPATH_FACTORY, null);
198        } catch (final XPathFactoryConfigurationException e) {
199            // newDefaultInstance declares no checked exception; mirror XPathFactory.newInstance(), which reports a default-model miss as a RuntimeException.
200            throw new RuntimeException("Neither XPathFactory.newDefaultInstance() nor " + JDK_XPATH_FACTORY + " is available", e);
201        }
202    }
203
204    /**
205     * Returns a new, secure {@link XPathFactory} for the default XPath object model.
206     *
207     * @return A secure factory.
208     * @throws IllegalStateException Thrown if a required secure setting cannot be applied to the underlying implementation.
209     * @throws RuntimeException      Thrown if there is a failure in creating an {@link XPathFactory} for the default object model.
210     */
211    public static XPathFactory newInstance() {
212        return secure(XPathFactory.newInstance());
213    }
214
215    /**
216     * Returns a new, secure {@link XPathFactory} for the given object model.
217     *
218     * @param uri The underlying object model identifier, as accepted by {@link XPathFactory#newInstance(String)}.
219     * @return A secure factory.
220     * @throws IllegalStateException              Thrown if a required secure setting cannot be applied to the underlying implementation.
221     * @throws XPathFactoryConfigurationException Thrown if no implementation of the object model is available.
222     * @throws NullPointerException               Thrown if {@code uri} is {@code null}.
223     * @throws IllegalArgumentException           Thrown if {@code uri} is empty.
224     */
225    public static XPathFactory newInstance(final String uri) throws XPathFactoryConfigurationException {
226        return secure(XPathFactory.newInstance(uri));
227    }
228
229    /**
230     * Returns a new, secure {@link XPathFactory} of the given implementation class.
231     *
232     * @param uri              The underlying object model identifier, as accepted by {@link XPathFactory#newInstance(String)}.
233     * @param factoryClassName The fully qualified class name of the {@link XPathFactory} implementation.
234     * @param classLoader      The class loader used to load the factory class; {@code null} means the current thread's context class loader.
235     * @return A secure factory.
236     * @throws IllegalStateException              Thrown if a required secure setting cannot be applied to the underlying implementation.
237     * @throws XPathFactoryConfigurationException Thrown if {@code factoryClassName} is {@code null}, or if the factory class cannot be loaded or
238     *                                            instantiated, or does not support {@code uri}.
239     * @throws NullPointerException               Thrown if {@code uri} is {@code null}.
240     * @throws IllegalArgumentException           Thrown if {@code uri} is empty.
241     */
242    public static XPathFactory newInstance(final String uri, final String factoryClassName, final ClassLoader classLoader)
243            throws XPathFactoryConfigurationException {
244        return secure(XPathFactory.newInstance(uri, factoryClassName, classLoader));
245    }
246
247    /**
248     * Applies capability-driven secure settings to any {@link XPathFactory} on the classpath.
249     *
250     * <p>
251     * The XPath object model mirrors TrAX: the stock JDK and Apache Xalan ship an XPath 1.0 engine with no URI-fetching functions, while Saxon adds the XPath
252     * 3.1
253     * {@code fn:doc}, {@code fn:collection} and {@code fn:unparsed-text} functions that can reach external resources. Rather than branching on the
254     * implementation class, this method probes what the factory supports and adapts:
255     * </p>
256     * <ul>
257     * <li><strong>Saxon</strong> ({@code net.sf.saxon}): recognized by package prefix and handed to {@link SaxonProvider#configure(XPathFactory)}, so any
258     * public         subclass routes to the same recipe as the registered factory. Its URI-fetching
259     * functions and reflection-based extension calls are reachable only through a locked-down Saxon {@code Configuration}, not the standard JAXP knobs; this
260     *         is the XPath counterpart of the Saxon exception in {@link SecureTransformerFactory#secure(javax.xml.transform.TransformerFactory)}, kept as a
261     *         documented package-prefix exception because the required securing surface is reachable only through a vendor API.</li>
262     *     <li><strong>FSP</strong> ({@link javax.xml.XMLConstants#FEATURE_SECURE_PROCESSING}): required. It is the only knob both the stock JDK and Xalan XPath
263     *         engines expose, and switches on their secure-processing limits. {@link XPathFactory} has no attribute API for finer control.</li>
264     *     <li><strong>The nested wrapper</strong>: required. FSP governs only the engine, not the parser it provisions internally for the
265     *         {@link org.xml.sax.InputSource}-taking {@code evaluate} entry points; the wrapper performs that document build with a secure parser instead, so
266     *         the engine never parses.</li>
267     * </ul>
268     *
269     * @param factory The factory to secure.
270     * @return A new secure factory or the original factory, as-is, if it is a known Saxon factory.
271     * @throws SecureException Thrown if this {@link XPathFactory} or the {@code XPath}s it creates cannot support this feature.
272     */
273    static XPathFactory secure(final XPathFactory factory) {
274        if (SaxonProvider.isSaxon(factory.getClass())) {
275            // Saxon: only a locked-down Configuration can close its URI-fetching functions and extension-function surface.
276            return SaxonProvider.configure(factory);
277        }
278        // Required: enables the engine's secure-processing limits; XPathFactory has no attribute API for finer control.
279        setFeature(factory, XMLConstants.FEATURE_SECURE_PROCESSING, true);
280        // Required: FSP does not reach the parser the engine provisions for InputSource-taking evaluate calls; the wrapper parses those itself.
281        return new Wrapper(factory);
282    }
283
284    /**
285     * Sets a feature on the given factory, throwing a {@link SecureException} if the implementation does not recognize it.
286     *
287     * @param factory The factory to secure.
288     * @param feature The feature to set.
289     * @param value   The value to set.
290     * @throws SecureException Thrown if this {@link XPathFactory} or the {@code XPath}s it creates cannot support this feature or if {@code feature} is
291     *                            {@code null}.
292     */
293    private static void setFeature(final XPathFactory factory, final String feature, final boolean value) {
294        try {
295            factory.setFeature(feature, value);
296        } catch (final XPathFactoryConfigurationException e) {
297            throw SecureException.featureFailed(feature, factory, e);
298        }
299    }
300
301    private SecureXPathFactory() {
302        // static only
303    }
304}