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 */
017package org.apache.commons.configuration2.io;
018
019import java.io.File;
020import java.net.MalformedURLException;
021import java.net.URI;
022import java.net.URL;
023import java.util.Arrays;
024import java.util.Map;
025
026import org.apache.commons.configuration2.ex.ConfigurationException;
027import org.apache.commons.lang3.ObjectUtils;
028import org.apache.commons.lang3.StringUtils;
029import org.apache.commons.logging.Log;
030import org.apache.commons.logging.LogFactory;
031
032/**
033 * <p>
034 * A utility class providing helper methods related to locating files.
035 * </p>
036 * <p>
037 * The methods of this class are used behind the scenes when retrieving configuration files based on different criteria,
038 * for example URLs, files, or more complex search strategies. They also implement functionality required by the default
039 * {@link FileSystem} implementations. Most methods are intended to be used internally only by other classes in the
040 * {@code io} package.
041 * </p>
042 *
043 * @since 2.0
044 */
045public final class FileLocatorUtils {
046
047    /**
048     * Constant for the default {@code FileSystem}. This file system is used by operations of this class if no specific file
049     * system is provided. An instance of {@link DefaultFileSystem} is used.
050     */
051    public static final FileSystem DEFAULT_FILE_SYSTEM = new DefaultFileSystem();
052
053    /**
054     * Constant for the default {@code FileLocationStrategy}. This strategy is used by the {@code locate()} method if the
055     * passed in {@code FileLocator} does not define its own location strategy. The default location strategy is roughly
056     * equivalent to the search algorithm used in version 1.x of <em>Commons Configuration</em> (there it was hard-coded
057     * though). It behaves in the following way when passed a {@code FileLocator}:
058     * <ul>
059     * <li>If the {@code FileLocator} has a defined URL, this URL is used as the file's URL (without any further
060     * checks).</li>
061     * <li>Otherwise, base path and file name stored in the {@code FileLocator} are passed to the current
062     * {@code FileSystem}'s {@code locateFromURL()} method. If this results in a URL, it is returned.</li>
063     * <li>Otherwise, if the locator's file name is an absolute path to an existing file, the URL of this file is
064     * returned.</li>
065     * <li>Otherwise, the concatenation of base path and file name is constructed. If this path points to an existing file,
066     * its URL is returned.</li>
067     * <li>Otherwise, a sub directory of the current user's home directory as defined by the base path is searched for the
068     * referenced file. If the file can be found there, its URL is returned.</li>
069     * <li>Otherwise, the base path is ignored, and the file name is searched in the current user's home directory. If the
070     * file can be found there, its URL is returned.</li>
071     * <li>Otherwise, a resource with the name of the locator's file name is searched in the classpath. If it can be found,
072     * its URL is returned.</li>
073     * <li>Otherwise, the strategy gives up and returns <strong>null</strong> indicating that the file cannot be resolved.</li>
074     * </ul>
075     */
076    // @formatter:off
077    public static final FileLocationStrategy DEFAULT_LOCATION_STRATEGY = newDefaultLocationStrategy();
078    // @formatter:on
079
080    /** Constant for the file URL protocol */
081    private static final String FILE_SCHEME = "file:";
082
083    /** The logger. */
084    private static final Log LOG = LogFactory.getLog(FileLocatorUtils.class);
085
086    /** Property key for the base path. */
087    private static final String PROP_BASE_PATH = "basePath";
088
089    /** Property key for the encoding. */
090    private static final String PROP_ENCODING = "encoding";
091
092    /** Property key for the file name. */
093    private static final String PROP_FILE_NAME = "fileName";
094
095    /** Property key for the file system. */
096    private static final String PROP_FILE_SYSTEM = "fileSystem";
097
098    /** Property key for the location strategy. */
099    private static final String PROP_STRATEGY = "locationStrategy";
100
101    /** Property key for the source URL. */
102    private static final String PROP_SOURCE_URL = "sourceURL";
103
104    /**
105     * Extends a path by another component. The given extension is added to the already existing path adding a separator if
106     * necessary.
107     *
108     * @param path the path to be extended
109     * @param ext the extension of the path
110     * @return the extended path
111     */
112    static String appendPath(final String path, final String ext) {
113        final StringBuilder fName = new StringBuilder();
114        fName.append(path);
115        // My best friend. Paranoia.
116        if (!path.endsWith(File.separator)) {
117            fName.append(File.separator);
118        }
119        //
120        // We have a relative path, and we have
121        // two possible forms here. If we have the
122        // "./" form then just strip that off first
123        // before continuing.
124        //
125        if (ext.startsWith("." + File.separator)) {
126            fName.append(ext.substring(2));
127        } else {
128            fName.append(ext);
129        }
130        return fName.toString();
131    }
132
133    /**
134     * Helper method for constructing a file object from a base path and a file name. This method is called if the base path
135     * passed to {@code getURL()} does not seem to be a valid URL.
136     *
137     * @param basePath the base path
138     * @param fileName the file name (must not be <strong>null</strong>)
139     * @return the resulting file
140     */
141    static File constructFile(final String basePath, final String fileName) {
142        final File file;
143        final File absolute = new File(fileName);
144        if (StringUtils.isEmpty(basePath) || absolute.isAbsolute()) {
145            file = absolute;
146        } else {
147            file = new File(appendPath(basePath, fileName));
148        }
149        return file;
150    }
151
152    /**
153     * Tries to convert the specified file to a URL. If this causes an exception, result is <strong>null</strong>.
154     *
155     * @param file the file to be converted
156     * @return the resulting URL or <strong>null</strong>
157     */
158    static URL convertFileToURL(final File file) {
159        return convertURIToURL(file.toURI());
160    }
161
162    /**
163     * Tries to convert the specified URI to a URL. If this causes an exception, result is <strong>null</strong>.
164     *
165     * @param uri the URI to be converted
166     * @return the resulting URL or <strong>null</strong>
167     */
168    static URL convertURIToURL(final URI uri) {
169        try {
170            return uri.toURL();
171        } catch (final MalformedURLException e) {
172            return null;
173        }
174    }
175
176    /**
177     * Creates a fully initialized {@code FileLocator} based on the specified URL.
178     *
179     * @param src the source {@code FileLocator}
180     * @param url the URL
181     * @return the fully initialized {@code FileLocator}
182     */
183    private static FileLocator createFullyInitializedLocatorFromURL(final FileLocator src, final URL url) {
184        final FileLocator.FileLocatorBuilder fileLocatorBuilder = fileLocator(src);
185        if (src.getSourceURL() == null) {
186            fileLocatorBuilder.sourceURL(url);
187        }
188        if (StringUtils.isBlank(src.getFileName())) {
189            fileLocatorBuilder.fileName(getFileName(url));
190        }
191        if (StringUtils.isBlank(src.getBasePath())) {
192            fileLocatorBuilder.basePath(getBasePath(url));
193        }
194        return fileLocatorBuilder.create();
195    }
196
197    /**
198     * Tries to convert the specified URL to a file object. If this fails, <strong>null</strong> is returned.
199     *
200     * @param url the URL
201     * @return the resulting file object
202     */
203    public static File fileFromURL(final URL url) {
204        return FileUtils.toFile(url);
205    }
206
207    /**
208     * Returns an uninitialized {@code FileLocatorBuilder} which can be used for the creation of a {@code FileLocator}
209     * object. This method provides a convenient way to create file locators using a fluent API as in the following example:
210     *
211     * <pre>
212     * FileLocator locator = FileLocatorUtils.fileLocator().basePath(myBasePath).fileName("test.xml").create();
213     * </pre>
214     *
215     * @return a builder object for defining a {@code FileLocator}
216     */
217    public static FileLocator.FileLocatorBuilder fileLocator() {
218        return fileLocator(null);
219    }
220
221    /**
222     * Returns a {@code FileLocatorBuilder} which is already initialized with the properties of the passed in
223     * {@code FileLocator}. This builder can be used to create a {@code FileLocator} object which shares properties of the
224     * original locator (for example the {@code FileSystem} or the encoding), but points to a different file. An example use case
225     * is as follows:
226     *
227     * <pre>
228     * FileLocator loc1 = ...
229     * FileLocator loc2 = FileLocatorUtils.fileLocator(loc1)
230     *     .setFileName("anotherTest.xml")
231     *     .create();
232     * </pre>
233     *
234     * @param src the source {@code FileLocator} (may be <strong>null</strong>)
235     * @return an initialized builder object for defining a {@code FileLocator}
236     */
237    public static FileLocator.FileLocatorBuilder fileLocator(final FileLocator src) {
238        return new FileLocator.FileLocatorBuilder(src);
239    }
240
241    /**
242     * Creates a new {@code FileLocator} object with the properties defined in the given map. The map must be conform to the
243     * structure generated by the {@link #put(FileLocator, Map)} method; unexpected data can cause
244     * {@code ClassCastException} exceptions. The map can be <strong>null</strong>, then an uninitialized {@code FileLocator} is
245     * returned.
246     *
247     * @param map the map
248     * @return the new {@code FileLocator}
249     * @throws ClassCastException if the map contains invalid data
250     */
251    public static FileLocator fromMap(final Map<String, ?> map) {
252        final FileLocator.FileLocatorBuilder builder = fileLocator();
253        if (map != null) {
254            builder.basePath((String) map.get(PROP_BASE_PATH)).encoding((String) map.get(PROP_ENCODING)).fileName((String) map.get(PROP_FILE_NAME))
255                .fileSystem((FileSystem) map.get(PROP_FILE_SYSTEM)).locationStrategy((FileLocationStrategy) map.get(PROP_STRATEGY))
256                .sourceURL((URL) map.get(PROP_SOURCE_URL));
257        }
258        return builder.create();
259    }
260
261    /**
262     * Returns a {@code FileLocator} object based on the passed in one whose location is fully defined. This method ensures
263     * that all components of the {@code FileLocator} pointing to the file are set in a consistent way. In detail it behaves
264     * as follows:
265     * <ul>
266     * <li>If the {@code FileLocator} has already all components set which define the file, it is returned unchanged.
267     * <em>Note:</em> It is not checked whether all components are really consistent!</li>
268     * <li>{@link #locate(FileLocator)} is called to determine a unique URL pointing to the referenced file. If this is
269     * successful, a new {@code FileLocator} is created as a copy of the passed in one, but with all components pointing to
270     * the file derived from this URL.</li>
271     * <li>Otherwise, result is <strong>null</strong>.</li>
272     * </ul>
273     *
274     * @param locator the {@code FileLocator} to be completed
275     * @return a {@code FileLocator} with a fully initialized location if possible or <strong>null</strong>
276     */
277    public static FileLocator fullyInitializedLocator(final FileLocator locator) {
278        if (isFullyInitialized(locator)) {
279            // already fully initialized
280            return locator;
281        }
282        final URL url = locate(locator);
283        return url != null ? createFullyInitializedLocatorFromURL(locator, url) : null;
284    }
285
286    /**
287     * Gets the path without the file name, for example https://xyz.net/foo/bar.xml results in https://xyz.net/foo/
288     *
289     * @param url the URL from which to extract the path
290     * @return the path component of the passed in URL
291     */
292    static String getBasePath(final URL url) {
293        if (url == null) {
294            return null;
295        }
296        String s = url.toString();
297        final String schemeHierPrefix = FILE_SCHEME + "//";
298        if (s.startsWith(FILE_SCHEME) && !s.startsWith(schemeHierPrefix)) {
299            s = schemeHierPrefix + s.substring(FILE_SCHEME.length());
300        }
301        if (s.endsWith("/") || StringUtils.isEmpty(url.getPath())) {
302            return s;
303        }
304        return s.substring(0, s.lastIndexOf("/") + 1);
305    }
306
307    /**
308     * Tries to find a resource with the given name in the classpath.
309     *
310     * @param resourceName the name of the resource
311     * @return the URL to the found resource or <strong>null</strong> if the resource cannot be found
312     */
313    static URL getClasspathResource(final String resourceName) {
314        URL url = null;
315        // attempt to load from the context classpath
316        final ClassLoader loader = Thread.currentThread().getContextClassLoader();
317        if (loader != null) {
318            url = loader.getResource(resourceName);
319
320            if (url != null) {
321                LOG.debug("Loading configuration from the context classpath (" + resourceName + ")");
322            }
323        }
324        // attempt to load from the system classpath
325        if (url == null) {
326            url = ClassLoader.getSystemResource(resourceName);
327
328            if (url != null) {
329                LOG.debug("Loading configuration from the system classpath (" + resourceName + ")");
330            }
331        }
332        return url;
333    }
334
335    /**
336     * Tries to convert the specified base path and file name into a file object. This method is called for example by the save()
337     * methods of file based configurations. The parameter strings can be relative files, absolute files and URLs as well.
338     * This implementation checks first whether the passed in file name is absolute. If this is the case, it is returned.
339     * Otherwise further checks are performed whether the base path and file name can be combined to a valid URL or a valid
340     * file name. <em>Note:</em> The test if the passed in file name is absolute is performed using
341     * {@code java.io.File.isAbsolute()}. If the file name starts with a slash, this method will return <strong>true</strong> on Unix,
342     * but <strong>false</strong> on Windows. So to ensure correct behavior for relative file names on all platforms you should never
343     * let relative paths start with a slash. E.g. in a configuration definition file do not use something like that:
344     *
345     * <pre>
346     * &lt;properties fileName="/subdir/my.properties"/&gt;
347     * </pre>
348     *
349     * Under Windows this path would be resolved relative to the configuration definition file. Under Unix this would be
350     * treated as an absolute path name.
351     *
352     * @param basePath the base path
353     * @param fileName the file name (must not be <strong>null</strong>)
354     * @return the file object (<strong>null</strong> if no file can be obtained)
355     */
356    static File getFile(final String basePath, final String fileName) {
357        // Check if the file name is absolute
358        final File f = new File(fileName);
359        if (f.isAbsolute()) {
360            return f;
361        }
362        // Check if URLs are involved
363        URL url;
364        try {
365            url = new URL(new URL(basePath), fileName);
366        } catch (final MalformedURLException mex1) {
367            try {
368                url = new URL(fileName);
369            } catch (final MalformedURLException mex2) {
370                url = null;
371            }
372        }
373        if (url != null) {
374            return fileFromURL(url);
375        }
376        return constructFile(basePath, fileName);
377    }
378
379    /**
380     * Extract the file name from the specified URL.
381     *
382     * @param url the URL from which to extract the file name
383     * @return the extracted file name
384     */
385    static String getFileName(final URL url) {
386        if (url == null) {
387            return null;
388        }
389        final String path = url.getPath();
390        if (path.endsWith("/") || StringUtils.isEmpty(path)) {
391            return null;
392        }
393        return path.substring(path.lastIndexOf("/") + 1);
394    }
395
396    /**
397     * Obtains a non-<strong>null</strong> {@code FileSystem} object from the passed in {@code FileLocator}. If the passed in
398     * {@code FileLocator} has a {@code FileSystem} object, it is returned. Otherwise, result is the default
399     * {@code FileSystem}.
400     *
401     * @param locator the {@code FileLocator} (may be <strong>null</strong>)
402     * @return the {@code FileSystem} to be used for this {@code FileLocator}
403     */
404    static FileSystem getFileSystem(final FileLocator locator) {
405        return locator != null ? ObjectUtils.getIfNull(locator.getFileSystem(), DEFAULT_FILE_SYSTEM) : DEFAULT_FILE_SYSTEM;
406    }
407
408    /**
409     * Gets a non <strong>null</strong> {@code FileLocationStrategy} object from the passed in {@code FileLocator}. If the
410     * {@code FileLocator} is not <strong>null</strong> and has a {@code FileLocationStrategy} defined, this strategy is returned.
411     * Otherwise, result is the default {@code FileLocationStrategy}.
412     *
413     * @param locator the {@code FileLocator}
414     * @return the {@code FileLocationStrategy} for this {@code FileLocator}
415     */
416    static FileLocationStrategy getLocationStrategy(final FileLocator locator) {
417        return locator != null ? ObjectUtils.getIfNull(locator.getLocationStrategy(), DEFAULT_LOCATION_STRATEGY) : DEFAULT_LOCATION_STRATEGY;
418    }
419
420    /**
421     * Returns a flag whether all components of the given {@code FileLocator} describing the referenced file are defined. In
422     * order to reference a file, it is not necessary that all components are filled in (for instance, the URL alone is
423     * sufficient). For some use cases however, it might be of interest to have different methods for accessing the
424     * referenced file. Also, depending on the filled out properties, there is a subtle difference how the file is accessed:
425     * If only the file name is set (and optionally the base path), each time the file is accessed a {@code locate()}
426     * operation has to be performed to uniquely identify the file. If however the URL is determined once based on the other
427     * components and stored in a fully defined {@code FileLocator}, it can be used directly to identify the file. If the
428     * passed in {@code FileLocator} is <strong>null</strong>, result is <strong>false</strong>.
429     *
430     * @param locator the {@code FileLocator} to be checked (may be <strong>null</strong>)
431     * @return a flag whether all components describing the referenced file are initialized
432     */
433    public static boolean isFullyInitialized(final FileLocator locator) {
434        if (locator == null) {
435            return false;
436        }
437        return locator.getBasePath() != null && locator.getFileName() != null && locator.getSourceURL() != null;
438    }
439
440    /**
441     * Checks whether the specified {@code FileLocator} contains enough information to locate a file. This is the case if a
442     * file name or a URL is defined. If the passed in {@code FileLocator} is <strong>null</strong>, result is <strong>false</strong>.
443     *
444     * @param locator the {@code FileLocator} to check
445     * @return a flag whether a file location is defined by this {@code FileLocator}
446     */
447    public static boolean isLocationDefined(final FileLocator locator) {
448        return locator != null && (locator.getFileName() != null || locator.getSourceURL() != null);
449    }
450
451    /**
452     * Locates the provided {@code FileLocator}, returning a URL for accessing the referenced file. This method uses a
453     * {@link FileLocationStrategy} to locate the file the passed in {@code FileLocator} points to. If the
454     * {@code FileLocator} contains itself a {@code FileLocationStrategy}, it is used. Otherwise, the default
455     * {@code FileLocationStrategy} is applied. The strategy is passed the locator and a {@code FileSystem}. The resulting
456     * URL is returned. If the {@code FileLocator} is <strong>null</strong>, result is <strong>null</strong>.
457     *
458     * @param locator the {@code FileLocator} to be resolved
459     * @return the URL pointing to the referenced file or <strong>null</strong> if the {@code FileLocator} could not be resolved
460     * @see #DEFAULT_LOCATION_STRATEGY
461     */
462    public static URL locate(final FileLocator locator) {
463        if (locator == null) {
464            return null;
465        }
466
467        return getLocationStrategy(locator).locate(getFileSystem(locator), locator);
468    }
469
470    /**
471     * Tries to locate the file referenced by the passed in {@code FileLocator}. If this fails, an exception is thrown. This
472     * method works like {@link #locate(FileLocator)}; however, in case of a failed location attempt an exception is thrown.
473     *
474     * @param locator the {@code FileLocator} to be resolved
475     * @return the URL pointing to the referenced file
476     * @throws ConfigurationException if the file cannot be resolved
477     */
478    public static URL locateOrThrow(final FileLocator locator) throws ConfigurationException {
479        final URL url = locate(locator);
480        if (url == null) {
481            throw new ConfigurationException("Could not locate: %s", locator);
482        }
483        return url;
484    }
485
486    /**
487     * Creates the default location strategy. This method creates a combined location strategy as described in the comment
488     * of the {@link #DEFAULT_LOCATION_STRATEGY} member field.
489     *
490     * @return the default {@code FileLocationStrategy}
491     * @since 2.15.0
492     */
493    public static FileLocationStrategy newDefaultLocationStrategy() {
494        // @formatter:off
495        return new CombinedLocationStrategy(Arrays.asList(
496               new ProvidedURLLocationStrategy(),
497               new FileSystemLocationStrategy(),
498               new AbsoluteNameLocationStrategy(),
499               new BasePathLocationStrategy(),
500               new HomeDirectoryLocationStrategy.Builder().setEvaluateBasePath(true).getUnchecked(),
501               new HomeDirectoryLocationStrategy.Builder().setEvaluateBasePath(false).getUnchecked(),
502               new ClasspathLocationStrategy()));
503        // @formatter:on
504    }
505
506    /**
507     * Stores the specified {@code FileLocator} in the given map. With the {@link #fromMap(Map)} method a new
508     * {@code FileLocator} with the same properties as the original one can be created.
509     *
510     * @param locator the {@code FileLocator} to be stored
511     * @param map the map in which to store the {@code FileLocator} (must not be <strong>null</strong>)
512     * @throws IllegalArgumentException if the map is <strong>null</strong>
513     */
514    public static void put(final FileLocator locator, final Map<String, Object> map) {
515        if (map == null) {
516            throw new IllegalArgumentException("Map must not be null.");
517        }
518        if (locator != null) {
519            map.put(PROP_BASE_PATH, locator.getBasePath());
520            map.put(PROP_ENCODING, locator.getEncoding());
521            map.put(PROP_FILE_NAME, locator.getFileName());
522            map.put(PROP_FILE_SYSTEM, locator.getFileSystem());
523            map.put(PROP_SOURCE_URL, locator.getSourceURL());
524            map.put(PROP_STRATEGY, locator.getLocationStrategy());
525        }
526    }
527
528    /**
529     * Convert the specified file into an URL. This method is equivalent to file.toURI().toURL(). It was used to work around
530     * a bug in the JDK preventing the transformation of a file into an URL if the file name contains a '#' character. See
531     * the issue CONFIGURATION-300 for more details. Now that we switched to JDK 1.4 we can directly use
532     * file.toURI().toURL().
533     *
534     * @param file the file to be converted into an URL
535     * @return a URL
536     * @throws MalformedURLException If the file protocol handler is not found (should not happen) or if an error occurred
537     *         while constructing the URL
538     */
539    static URL toURL(final File file) throws MalformedURLException {
540        return file.toURI().toURL();
541    }
542
543    /**
544     * Private constructor so that no instances can be created.
545     */
546    private FileLocatorUtils() {
547    }
548
549}