sanitize more paths on windows, fixes #2335

Author: tarzanek

opengrok-indexer/src/main/java/org/opengrok/indexer/analysis/AnalyzerGuru.java

@@ -523,12 +523,9 @@ public void populateDocument(Document doc, File file, String path,
  FileAnalyzer fa, Writer xrefOut) throws IOException,
  InterruptedException, ForbiddenSymlinkException {
 
- // Sanitize Windows path delimiters in order not to conflict with Lucene escape character
- // and also so the path appears as correctly formed URI in the search results.
- path = path.replace("\\", "/");
-
  String date = DateTools.timeToString(file.lastModified(),
  DateTools.Resolution.MILLISECOND);
+ path = Util.fixPathIfWindows(path);
  doc.add(new Field(QueryBuilder.U, Util.path2uid(path, date),
  string_ft_stored_nanalyzed_norms));
  doc.add(new Field(QueryBuilder.FULLPATH, file.getAbsolutePath(),

opengrok-indexer/src/main/java/org/opengrok/indexer/configuration/Project.java

@@ -34,6 +34,7 @@
 import org.opengrok.indexer.logger.LoggerFactory;
 import org.opengrok.indexer.util.ClassUtil;
 import org.opengrok.indexer.util.ForbiddenSymlinkException;
+import org.opengrok.indexer.web.Util;
 
 /**
  * Placeholder for the information that builds up a project
@@ -112,7 +113,7 @@ public Project(String name) {
  */
  public Project(String name, String path) {
  this.name = name;
- this.path = path;
+ this.path = Util.fixPathIfWindows(path);
  completeWithDefaults();
  }
 
@@ -332,7 +333,7 @@ public static Project getProject(String path) {
  // Try to match each project path as prefix of the given path.
  final RuntimeEnvironment env = RuntimeEnvironment.getInstance();
  if (env.hasProjects()) {
- final String lpath = path.replace(File.separatorChar, '/');
+ final String lpath = Util.fixPathIfWindows(path);
  for (Project p : env.getProjectList()) {
  String projectPath = p.getPath();
  if (projectPath == null) {

opengrok-indexer/src/main/java/org/opengrok/indexer/index/IndexDatabase.java

@@ -464,6 +464,8 @@ public void update(IndexerParallelizer parallelizer)
  }
  }
 
+ dir = Util.fixPathIfWindows(dir);
+
  String startuid = Util.path2uid(dir, "");
  reader = DirectoryReader.open(indexDirectory); // open existing index
  settings = readAnalysisSettings();
@@ -1100,6 +1102,7 @@ private void indexDown(File dir, String parent, IndexDownArgs args)
  }
 
  if (uidIter != null) {
+ path = Util.fixPathIfWindows(path);
  String uid = Util.path2uid(path,
  DateTools.timeToString(file.lastModified(),
  DateTools.Resolution.MILLISECOND)); // construct uid for doc

opengrok-indexer/src/main/java/org/opengrok/indexer/index/Indexer.java

@@ -71,6 +71,12 @@
 /**
  * Creates and updates an inverted source index as well as generates Xref, file
  * stats etc., if specified in the options
+ *
+ * We shall use / as path delimiter in whole opengrok for uuids and paths
+ * from Windows systems, the path shall be converted when entering the index or web
+ * and converted back if needed* to access original file
+ *
+ * *Windows already supports opening /var/opengrok as C:\var\opengrok
  */
 @SuppressWarnings({"PMD.AvoidPrintStackTrace", "PMD.SystemPrintln"})
 public final class Indexer {
@@ -83,6 +89,10 @@ public final class Indexer {
  private static final String DIRBASED = "dirbased";
  private static final String UIONLY = "uionly";
 
+ //whole app uses this separator
+ public static final char PATH_SEPARATOR ='/';
+ public static String PATH_SEPARATOR_STRING =Character.toString(PATH_SEPARATOR);
+
  private static final Indexer index = new Indexer();
  private static Configuration cfg = null;
  private static boolean checkIndexVersion = false;

opengrok-indexer/src/main/java/org/opengrok/indexer/search/context/OGKUnifiedHighlighter.java

@@ -297,6 +297,7 @@ private String getRepoFileContent(String repoRelPath, String storedU)
  return null;
  }
 
+ repoRelPath = Util.fixPathIfWindows(repoRelPath);
  // Verify that timestamp (U) is unchanged by comparing UID.
  String uid = Util.path2uid(repoRelPath,
  DateTools.timeToString(repoAbsFile.lastModified(),

opengrok-indexer/src/main/java/org/opengrok/indexer/web/PageConfig.java

@@ -75,6 +75,9 @@
 import org.suigeneris.jrcs.diff.Diff;
 import org.suigeneris.jrcs.diff.DifferentiationFailedException;
 
+import static org.opengrok.indexer.index.Indexer.PATH_SEPARATOR;
+import static org.opengrok.indexer.index.Indexer.PATH_SEPARATOR_STRING;
+
 /**
  * A simple container to lazy initialize common vars wrt. a single request. It
  * MUST NOT be shared between several requests and
@@ -203,7 +206,7 @@ public String getHeaderData() {
  */
  public DiffData getDiffData() {
  DiffData data = new DiffData();
- data.path = getPath().substring(0, path.lastIndexOf('/'));
+ data.path = getPath().substring(0, path.lastIndexOf(PATH_SEPARATOR));
  data.filename = Util.htmlize(getResourceFile().getName());
 
  String srcRoot = getSourceRootPath();
@@ -736,7 +739,7 @@ public String getCrossFilename() {
  * option.
  *
  * @return always an array of 3 fields, whereby field[0] contains the path
- * value to use (starts and ends always with a '/'). Field[1] the contains
+ * value to use (starts and ends always with a {@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR}). Field[1] the contains
  * string to show in the UI. field[2] is set to {@code disabled=""} if the
  * current path is the "/" directory, otherwise set to an empty string.
  */
@@ -747,7 +750,7 @@ public String[] getSearchOnlyIn() {
  : new String[]{path, "this directory", ""};
  }
  String[] res = new String[3];
- res[0] = path.substring(0, path.lastIndexOf('/') + 1);
+ res[0] = path.substring(0, path.lastIndexOf(PATH_SEPARATOR) + 1);
  res[1] = res[0];
  res[2] = "";
  return res;
@@ -971,7 +974,7 @@ public String getTitle() {
  * @see RuntimeEnvironment#getWebappLAF()
  */
  public String getCssDir() {
- return req.getContextPath() + '/' + getEnv().getWebappLAF();
+ return req.getContextPath() + PATH_SEPARATOR + getEnv().getWebappLAF();
  }
 
  /**
@@ -1002,7 +1005,7 @@ public IgnoredNames getIgnoredNames() {
 
  /**
  * Get the canonical path to root of the source tree. File separators are
- * replaced with a '/'.
+ * replaced with a {@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR}.
  *
  * @return The on disk source root directory.
  * @see RuntimeEnvironment#getSourceRootPath()
@@ -1011,7 +1014,7 @@ public String getSourceRootPath() {
  if (sourceRootPath == null) {
  String srcpath = getEnv().getSourceRootPath();
  if (srcpath != null) {
- sourceRootPath = srcpath.replace(File.separatorChar, '/');
+ sourceRootPath = srcpath.replace(File.separatorChar, PATH_SEPARATOR);
  }
  }
  return sourceRootPath;
@@ -1032,7 +1035,7 @@ public Prefix getPrefix() {
 
  /**
  * Get the canonical path of the related resource relative to the source
- * root directory (used file separators are all '/'). No check is made,
+ * root directory (used file separators are all {@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR}). No check is made,
  * whether the obtained path is really an accessible resource on disk.
  *
  * @see HttpServletRequest#getPathInfo()
@@ -1041,8 +1044,8 @@ public Prefix getPrefix() {
  */
  public String getPath() {
  if (path == null) {
- path = Util.getCanonicalPath(req.getPathInfo(), '/');
- if ("/".equals(path)) {
+ path = Util.getCanonicalPath(req.getPathInfo(), PATH_SEPARATOR);
+ if (PATH_SEPARATOR_STRING.equals(path)) {
  path = "";
  }
  }
@@ -1076,7 +1079,7 @@ public File getResourceFile(String path) {
  * NOTE: If a repository contains hard or symbolic links, the returned file
  * may finally point to a file outside of the source root directory.
  *
- * @return {@code new File("/")} if the related file or directory is not
+ * @return {@code new File({@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR_STRING })} if the related file or directory is not
  * available (can not be find below the source root directory), the readable
  * file or directory otherwise.
  * @see #getSourceRootPath()
@@ -1086,23 +1089,23 @@ public File getResourceFile() {
  if (resourceFile == null) {
  resourceFile = getResourceFile(getPath());
  if (resourceFile == null) {
- resourceFile = new File("/");
+ resourceFile = new File(PATH_SEPARATOR_STRING);
  }
  }
  return resourceFile;
  }
 
  /**
  * Get the canonical on disk path to the request related file or directory
- * with all file separators replaced by a '/'.
+ * with all file separators replaced by a {@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR}.
  *
- * @return "/" if the evaluated path is invalid or outside the source root
+ * @return {@link org.opengrok.indexer.index.Indexer#PATH_SEPARATOR_STRING} if the evaluated path is invalid or outside the source root
  * directory), otherwise the path to the readable file or directory.
  * @see #getResourceFile()
  */
  public String getResourcePath() {
  if (resourcePath == null) {
- resourcePath = getResourceFile().getPath().replace(File.separatorChar, '/');
+ resourcePath = Util.fixPathIfWindows(getResourceFile().getPath());
  }
  return resourcePath;
  }
@@ -1119,7 +1122,7 @@ public String getResourcePath() {
  */
  public boolean resourceNotAvailable() {
  getIgnoredNames();
- return getResourcePath().equals("/") || ignoredNames.ignore(getPath())
+ return getResourcePath().equals(PATH_SEPARATOR_STRING) || ignoredNames.ignore(getPath())
  || ignoredNames.ignore(resourceFile.getParentFile())
  || ignoredNames.ignore(resourceFile);
  }
@@ -1137,8 +1140,8 @@ public boolean isDir() {
  }
 
  private static String trailingSlash(String path) {
- return path.length() == 0 || path.charAt(path.length() - 1) != '/'
- ? "/"
+ return path.length() == 0 || path.charAt(path.length() - 1) != PATH_SEPARATOR
+ ? PATH_SEPARATOR_STRING
  : "";
  }
 
@@ -1162,7 +1165,7 @@ private File checkFile(File dir, String name, boolean compressed) {
  private File checkFileResolve(File dir, String name, boolean compressed) {
  File lresourceFile = new File(getSourceRootPath() + getPath(), name);
  if (!lresourceFile.canRead()) {
- lresourceFile = new File("/");
+ lresourceFile = new File(PATH_SEPARATOR_STRING);
  }
  File f;
  if (compressed) {

opengrok-indexer/src/main/java/org/opengrok/indexer/web/Util.java

@@ -75,6 +75,8 @@
 import org.opengrok.indexer.logger.LoggerFactory;
 import org.opengrok.indexer.web.messages.MessagesContainer.AcceptedMessage;
 
+import static org.opengrok.indexer.index.Indexer.PATH_SEPARATOR;
+
 /**
  * Class for useful functions.
  */
@@ -96,6 +98,13 @@ public final class Util {
  private final static Pattern NON_ASCII_ALPHA_NUM = Pattern.compile(
  "[^A-Za-z0-9_]");
 
+ private static String OS = null;
+
+ private static final String anchorLinkStart = "<a href=\"";
+ private static final String anchorClassStart = "<a class=\"";
+ private static final String anchorEnd = "</a>";
+ private static final String closeQuotedTag = "\">";
+
  private Util() {
  // private to ensure static
  }
@@ -377,7 +386,7 @@ public static String versionParameter() {
  }
 
  /**
- * Convenience method for {@code breadcrumbPath(urlPrefix, path, '/')}.
+ * Convenience method for {@code breadcrumbPath(urlPrefix, path, PATH_SEPARATOR)}.
  *
  * @param urlPrefix prefix to add to each url
  * @param path path to crack
@@ -386,14 +395,9 @@ public static String versionParameter() {
  * @see #breadcrumbPath(String, String, char)
  */
  public static String breadcrumbPath(String urlPrefix, String path) {
- return breadcrumbPath(urlPrefix, path, '/');
+ return breadcrumbPath(urlPrefix, path, PATH_SEPARATOR);
  }
 
- private static final String anchorLinkStart = "<a href=\"";
- private static final String anchorClassStart = "<a class=\"";
- private static final String anchorEnd = "</a>";
- private static final String closeQuotedTag = "\">";
-
  /**
  * Convenience method for
  * {@code breadcrumbPath(urlPrefix, path, sep, "", false)}.
@@ -472,13 +476,13 @@ public static String breadcrumbPath(String urlPrefix, String path,
  * (17 + prefix.length() + postfix.length()));
  int k = path.indexOf(pnames[0]);
  if (path.lastIndexOf(sep, k) != -1) {
- pwd.append('/');
+ pwd.append(PATH_SEPARATOR);
  markup.append(sep);
  }
  for (int i = 0; i < pnames.length; i++) {
  pwd.append(URIEncodePath(pnames[i]));
  if (isDir || i < pnames.length - 1) {
- pwd.append('/');
+ pwd.append(PATH_SEPARATOR);
  }
  markup.append(anchorLinkStart).append(prefix).append(pwd)
  .append(postfix).append(closeQuotedTag).append(pnames[i])
@@ -855,8 +859,29 @@ public static String path2uid(String path, String date) {
  * @return the original path.
  */
  public static String uid2url(String uid) {
- String url = uid.replace('\u0000', '/');
- return url.substring(0, url.lastIndexOf('/')); // remove date from end
+ String url = uid.replace('\u0000', PATH_SEPARATOR);
+ return url.substring(0, url.lastIndexOf(PATH_SEPARATOR)); // remove date from end
+ }
+
+ public static String fixPathIfWindows(String path) {
+ if (Util.isWindows()) {
+ // Sanitize Windows path delimiters in order not to conflict with Lucene escape character
+ // and also so the path appears as correctly formed URI in the search results.
+ return path.replace(File.separatorChar, PATH_SEPARATOR);
+ }
+ return path;
+ }
+
+ public static String getOsName() {
+ if (OS == null) {
+ OS = System.getProperty("os.name");
+ }
+ return OS;
+ }
+
+ public static boolean isWindows() {
+ String osname = getOsName();
+ return osname != null ? osname.startsWith("Windows") : false;
  }
 
  /**

opengrok-indexer/src/test/java/org/opengrok/indexer/web/UtilTest.java

@@ -155,6 +155,14 @@ public void path2uid() {
  Util.path2uid("/etc/passwd", "date"));
  }
 
+ @Test
+ public void fixPathIfWindows() {
+ if (Util.isWindows()) {
+ assertEquals("/var/opengrok",
+ Util.fixPathIfWindows("\\var\\opengrok"));
+ }
+ }
+
  @Test
  public void uid2url() {
  assertEquals("/etc/passwd", Util.uid2url(