QUrl: add NormalizePathSegments to UrlFormattingOptions

This is a bit like QDir::cleanPath(), but for URL paths.
The code is shared with QDir::cleanPath(), by extracting the common parts
it into a helper, qt_normalizePathSegments().

Change-Id: I7133c5e4aa2bf17fba98af13eb5371afba64197a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

src/corelib/io/qdir.cpp

@@ -2005,25 +2005,14 @@ bool QDir::match(const QString &filter, const QString &fileName)
 #endif // QT_NO_REGEXP
 
 /*!
-    Returns \a path with directory separators normalized (converted to "/") and
-    redundant ones removed, and "."s and ".."s resolved (as far as possible).
+    Returns \a path with redundant directory separators removed,
+    and "."s and ".."s resolved (as far as possible).
 
-    Symbolic links are kept. This function does not return the
-    canonical path, but rather the simplest version of the input.
-    For example, "./local" becomes "local", "local/../bin" becomes
-    "bin" and "/local/usr/../bin" becomes "/local/bin".
-
-    \sa absolutePath(), canonicalPath()
+    This method is shared with QUrl, so it doesn't deal with QDir::separator(),
+    nor does it remove the trailing slash, if any.
 */
-QString QDir::cleanPath(const QString &path)
+QString qt_normalizePathSegments(const QString &name, bool allowUncPaths)
 {
-    if (path.isEmpty())
-        return path;
-    QString name = path;
-    QChar dir_separator = separator();
-    if (dir_separator != QLatin1Char('/'))
-       name.replace(dir_separator, QLatin1Char('/'));
-
     int used = 0, levels = 0;
     const int len = name.length();
     QVarLengthArray<QChar> outVector(len);
@@ -2033,10 +2022,8 @@ QString QDir::cleanPath(const QString &path)
     for (int i = 0, last = -1, iwrite = 0; i < len; ++i) {
         if (p[i] == QLatin1Char('/')) {
             while (i+1 < len && p[i+1] == QLatin1Char('/')) {
-#if defined(Q_OS_WIN) && !defined(Q_OS_WINCE) //allow unc paths
-                if (!i)
+                if (allowUncPaths && i == 0)
                     break;
-#endif
                 i++;
             }
             bool eaten = false;
@@ -2099,8 +2086,6 @@ QString QDir::cleanPath(const QString &path)
                 eaten = true;
 #endif
                 last = -1;
-            } else if (last != -1 && i == len-1) {
-                eaten = true;
             } else {
                 levels++;
             }
@@ -2126,6 +2111,36 @@ QString QDir::cleanPath(const QString &path)
     }
 
     QString ret = (used == len ? name : QString(out, used));
+    return ret;
+}
+
+/*!
+    Returns \a path with directory separators normalized (converted to "/") and
+    redundant ones removed, and "."s and ".."s resolved (as far as possible).
+
+    Symbolic links are kept. This function does not return the
+    canonical path, but rather the simplest version of the input.
+    For example, "./local" becomes "local", "local/../bin" becomes
+    "bin" and "/local/usr/../bin" becomes "/local/bin".
+
+    \sa absolutePath(), canonicalPath()
+*/
+QString QDir::cleanPath(const QString &path)
+{
+    if (path.isEmpty())
+        return path;
+    QString name = path;
+    QChar dir_separator = separator();
+    if (dir_separator != QLatin1Char('/'))
+       name.replace(dir_separator, QLatin1Char('/'));
+
+    bool allowUncPaths = false;
+#if defined(Q_OS_WIN) && !defined(Q_OS_WINCE) //allow unc paths
+    allowUncPaths = true;
+#endif
+
+    QString ret = qt_normalizePathSegments(name, allowUncPaths);
+
     // Strip away last slash except for root directories
     if (ret.length() > 1 && ret.endsWith(QLatin1Char('/'))) {
 #if defined (Q_OS_WIN)

src/corelib/io/qurl.cpp

@@ -231,6 +231,8 @@
     \value PreferLocalFile If the URL is a local file according to isLocalFile()
      and contains no query or fragment, a local file path is returned.
     \value StripTrailingSlash  The trailing slash is removed if one is present.
+    \value NormalizePathSegments  Modifies the path to remove redundant directory separators,
+             and to resolve "."s and ".."s (as far as possible).
 
     Note that the case folding rules in \l{RFC 3491}{Nameprep}, which QUrl
     conforms to, require host names to always be converted to lower case,
@@ -324,6 +326,7 @@
 #endif
 
 QT_BEGIN_NAMESPACE
+extern QString qt_normalizePathSegments(const QString &name, bool allowUncPaths); // qdir.cpp
 
 inline static bool isHex(char c)
 {
@@ -803,6 +806,9 @@ inline void QUrlPrivate::appendPassword(QString &appendTo, QUrl::FormattingOptio
 inline void QUrlPrivate::appendPath(QString &appendTo, QUrl::FormattingOptions options, Section appendingTo) const
 {
     QString thePath = path;
+    if (options & QUrl::NormalizePathSegments) {
+        thePath = qt_normalizePathSegments(path, false);
+    }
     if (options & QUrl::RemoveFilename) {
         const int slash = path.lastIndexOf(QLatin1Char('/'));
         if (slash == -1)
@@ -3235,7 +3241,7 @@ QUrl QUrl::adjusted(QUrl::FormattingOptions options) const
         that.setFragment(QString());
     if (options & RemovePath) {
         that.setPath(QString());
-    } else if (options & (StripTrailingSlash | RemoveFilename)) {
+    } else if (options & (StripTrailingSlash | RemoveFilename | NormalizePathSegments)) {
         QString path;
         d->appendPath(path, options, QUrlPrivate::Path);
         that.setPath(path);

src/corelib/io/qurl.h

@@ -141,7 +141,8 @@ public:
         // 0x100 was a private code in Qt 4, keep unused for a while
         PreferLocalFile = 0x200,
         StripTrailingSlash = 0x400,
-        RemoveFilename = 0x800
+        RemoveFilename = 0x800,
+        NormalizePathSegments = 0x1000
     };
 
     enum ComponentFormattingOption {

tests/auto/corelib/io/qurl/tst_qurl.cpp

@@ -298,6 +298,17 @@ void tst_QUrl::comparison()
     QEXPECT_FAIL("", "Normalization not implemented, will probably not be implemented like this", Continue);
     QCOMPARE(url3, url4);
 
+    QUrl url3bis = QUrl::fromEncoded("example://a/b/c/%7Bfoo%7D/");
+    QUrl url3bisNoSlash = QUrl::fromEncoded("example://a/b/c/%7Bfoo%7D");
+    QUrl url4bis = QUrl::fromEncoded("example://a/.//b/../b/c//%7Bfoo%7D/");
+    QCOMPARE(url4bis.adjusted(QUrl::NormalizePathSegments), url3bis);
+    QCOMPARE(url4bis.adjusted(QUrl::NormalizePathSegments | QUrl::StripTrailingSlash), url3bisNoSlash);
+
+    QUrl url4EncodedDots = QUrl("example://a/.//b/%2E%2E%2F/b/c/");
+    QCOMPARE(QString::fromLatin1(url4EncodedDots.toEncoded()), QString::fromLatin1("example://a/.//b/..%2F/b/c/"));
+    QCOMPARE(url4EncodedDots.toString(), QString("example://a/.//b/..%2F/b/c/"));
+    QCOMPARE(url4EncodedDots.adjusted(QUrl::NormalizePathSegments).toString(), QString("example://a/b/..%2F/b/c/"));
+
     // 6.2.2.1 Make sure hexdecimal characters in percent encoding are
     // treated case-insensitively
     QUrl url5;