Make url normalization closer to common browser behavior

Firefox, Chrome and various http libraries normalize /./ and /../ from
urls, but retain multiple adjacent slashes as is. Qt removes
duplicated slashes which makes it impossible to access some web
resources that rely on those.

Fixes: QTBUG-71973
Change-Id: Ie18ae6ad3264acb252fcd87a754726a8c546e5ec
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

1 files changed, 54 insertions, 10 deletions

diff --git a/src/corelib/io/qdir.cpp b/src/corelib/io/qdir.cpp
index 7df461ddce..405718aba8 100644
--- a/src/corelib/io/qdir.cpp
+++ b/src/corelib/io/qdir.cpp
@@ -2161,9 +2161,10 @@ bool QDir::match(const QString &filter, const QString &fileName)
     This method is shared with QUrl, so it doesn't deal with QDir::separator(),
     nor does it remove the trailing slash, if any.
 */
-Q_AUTOTEST_EXPORT QString qt_normalizePathSegments(const QString &name, bool allowUncPaths,
-                                                   bool *ok = nullptr)
+QString qt_normalizePathSegments(const QString &name, QDirPrivate::PathNormalizations flags, bool *ok)
 {
+    const bool allowUncPaths = QDirPrivate::AllowUncPaths & flags;
+    const bool isRemote = QDirPrivate::RemotePath & flags;
     const int len = name.length();
 
     if (ok)
@@ -2185,14 +2186,30 @@ Q_AUTOTEST_EXPORT QString qt_normalizePathSegments(const QString &name, bool all
     i -= prefixLength;
 
     // replicate trailing slash (i > 0 checks for emptiness of input string p)
-    if (i > 0 && p[i] == '/') {
+    // except for remote paths because there can be /../ or /./ ending
+    if (i > 0 && p[i] == '/' && !isRemote) {
         out[--used] = '/';
         --i;
     }
 
+    auto isDot = [](const ushort *p, int i) {
+        return i > 1 && p[i - 1] == '.' && p[i - 2] == '/';
+    };
+    auto isDotDot = [](const ushort *p, int i) {
+        return i > 2 && p[i - 1] == '.' && p[i - 2] == '.' && p[i - 3] == '/';
+    };
+
     while (i >= 0) {
-        // remove trailing slashes
+        // copy trailing slashes for remote urls
         if (p[i] == '/') {
+            if (isRemote && !up) {
+                if (isDot(p, i)) {
+                    i -= 2;
+                    continue;
+                }
+                out[--used] = p[i];
+            }
+
             --i;
             continue;
         }
@@ -2204,10 +2221,17 @@ Q_AUTOTEST_EXPORT QString qt_normalizePathSegments(const QString &name, bool all
         }
 
         // detect up dir
-        if (i >= 1 && p[i] == '.' && p[i-1] == '.'
-                && (i == 1 || (i >= 2 && p[i-2] == '/'))) {
+        if (i >= 1 && p[i] == '.' && p[i-1] == '.' && (i < 2 || p[i - 2] == '/')) {
             ++up;
-            i -= 2;
+            i -= i >= 2 ? 3 : 2;
+
+            if (isRemote) {
+                // moving up should consider empty path segments too (/path//../ -> /path/)
+                while (i > 0 && up && p[i] == '/') {
+                    --up;
+                    --i;
+                }
+            }
             continue;
         }
 
@@ -2217,7 +2241,27 @@ Q_AUTOTEST_EXPORT QString qt_normalizePathSegments(const QString &name, bool all
 
         // skip or copy
         while (i >= 0) {
-            if (p[i] == '/') { // do not copy slashes
+            if (p[i] == '/') {
+                // copy all slashes as is for remote urls if they are not part of /./ or /../
+                if (isRemote && !up) {
+                    while (i > 0 && p[i] == '/' && !isDotDot(p, i)) {
+
+                        if (isDot(p, i)) {
+                            i -= 2;
+                            continue;
+                        }
+
+                        out[--used] = p[i];
+                        --i;
+                    }
+
+                    // in case of /./, jump over
+                    if (isDot(p, i))
+                        i -= 2;
+
+                    break;
+                }
+
                 --i;
                 break;
             }
@@ -2238,7 +2282,7 @@ Q_AUTOTEST_EXPORT QString qt_normalizePathSegments(const QString &name, bool all
         *ok = prefixLength == 0 || up == 0;
 
     // add remaining '..'
-    while (up) {
+    while (up && !isRemote) {
         if (used != len && out[used] != '/') // is not empty and there isn't already a '/'
             out[--used] = '/';
         out[--used] = '.';
@@ -2284,7 +2328,7 @@ static QString qt_cleanPath(const QString &path, bool *ok)
     if (dir_separator != QLatin1Char('/'))
        name.replace(dir_separator, QLatin1Char('/'));
 
-    QString ret = qt_normalizePathSegments(name, OSSupportsUncPaths, ok);
+    QString ret = qt_normalizePathSegments(name, OSSupportsUncPaths ? QDirPrivate::AllowUncPaths : QDirPrivate::DefaultNormalization, ok);
 
     // Strip away last slash except for root directories
     if (ret.length() > 1 && ret.endsWith(QLatin1Char('/'))) {