QUrl: expand the square brackets encoding to decoded setPath() calls

Amends (reverts) commit 5e936b60fc921e21b8153a83113886a1de333b57, which
forced the encoding of the square brackets ("[]") into their percent-
encoded forms only in QUrl::fromLocalFile(). This commit expands the
functionality to all uses of decoded paths by applying the change
directly to recodeFromUser().

[ChangeLog][QtCore][QUrl] Square brackets ("[]") are now transformed to
their percent-encoded forms ("%5B" and "%5D") when present as inputs to
setPath(), setQuery(), and setFragment() if the parsing mode is
QUrl::DecodedMode (the default).

Pick-to: 6.8
Fixes: QTBUG-135433
Task-number: QTBUG-134073
See: https://bugs.kde.org/show_bug.cgi?id=502280
Change-Id: Id2b41559af08b5f228e4fffdb9c6f23120f856b5
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit f3da9d3c858e8dc28a5dc91047b592ed5becbd62)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>

src/corelib/io/qurl.cpp

@@ -748,32 +748,6 @@ static const ushort * const pathInIsolation = userNameInIsolation + 5;
 static const ushort * const queryInIsolation = userNameInIsolation + 6;
 static const ushort * const fragmentInIsolation = userNameInIsolation + 7;
 
-static const ushort localPathFromUser[] = {
-    // we force-decode some of the gen-delims, because
-    //    pchar         = unreserved / pct-encoded / sub-delims / ":" / "@"
-    // the gen-delim lines are leave() in qt_urlRecode, so we don't need to
-    // repeat them if we want to keep them decoded
-    // decode(':'), // allowed
-    // decode('@'), // allowed
-    encode(']'),
-    encode('['),
-    // decode('/'), // special and allowed
-    // decode('?'), // handled by path() and others
-    // decode('#'), // ditto
-
-    // the rest is like pathInIsolation above
-    decode('"'),
-    decode('<'),
-    decode('>'),
-    decode('^'),
-    decode('\\'),
-    decode('|'),
-    decode('{'),
-    decode('}'),
-
-    0
-};
-
 static const ushort userNameInUserInfo[] =  {
     encode(':'), // 0
     decode('@'), // 1
@@ -833,9 +807,11 @@ static const ushort * const pathInUrl = userNameInUrl + 5;
 static const ushort * const queryInUrl = userNameInUrl + 6;
 static const ushort * const fragmentInUrl = userNameInUrl + 6;
 
-static inline void parseDecodedComponent(QString &data)
+static inline void parseDecodedComponent(QString &data, QUrlPrivate::Section section)
 {
     data.replace(u'%', "%25"_L1);
+    if (section != QUrlPrivate::Host)
+        data.replace(u'[', "%5B"_L1).replace(u']', "%5D"_L1);
 }
 
 static inline QString
@@ -2135,7 +2111,7 @@ void QUrl::setUserName(const QString &userName, ParsingMode mode)
 
     QString data = userName;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::UserName);
         mode = TolerantMode;
     }
 
@@ -2198,7 +2174,7 @@ void QUrl::setPassword(const QString &password, ParsingMode mode)
 
     QString data = password;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::Password);
         mode = TolerantMode;
     }
 
@@ -2260,7 +2236,7 @@ void QUrl::setHost(const QString &host, ParsingMode mode)
 
     QString data = host;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::Host);
         mode = TolerantMode;
     }
 
@@ -2385,7 +2361,7 @@ void QUrl::setPath(const QString &path, ParsingMode mode)
 
     QString data = path;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::Path);
         mode = TolerantMode;
     }
 
@@ -2521,7 +2497,7 @@ void QUrl::setQuery(const QString &query, ParsingMode mode)
 
     QString data = query;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::Query);
         mode = TolerantMode;
     }
 
@@ -2619,7 +2595,7 @@ void QUrl::setFragment(const QString &fragment, ParsingMode mode)
 
     QString data = fragment;
     if (mode == DecodedMode) {
-        parseDecodedComponent(data);
+        parseDecodedComponent(data, QUrlPrivate::Fragment);
         mode = TolerantMode;
     }
 
@@ -3386,11 +3362,7 @@ QUrl QUrl::fromLocalFile(const QString &localFile)
     }
 
     url.setScheme(scheme);
-
+    url.setPath(deslashified, DecodedMode);
-    // not directly using setPath here, as we do a few more transforms
-    parseDecodedComponent(deslashified);
-    if (!qt_urlRecode(url.d->path, deslashified, {}, localPathFromUser))
-        url.d->path = std::move(deslashified);
 
     return url;
 }

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

@@ -1422,6 +1422,11 @@ void tst_QUrl::toLocalFile()
     url.setPath(url.path(QUrl::PrettyDecoded), QUrl::TolerantMode);
     QCOMPARE(url.toLocalFile(), theFile);
     QCOMPARE(url.isLocalFile(), !theFile.isEmpty());
+
+    // local file paths can be fully decoded without loss
+    url.setPath(url.path());
+    QCOMPARE(url.toLocalFile(), theFile);
+    QCOMPARE(url.isLocalFile(), !theFile.isEmpty());
 }
 
 void tst_QUrl::fromLocalFile_data()
@@ -1508,6 +1513,11 @@ void tst_QUrl::fromLocalFile()
     url.setPath(url.path(QUrl::PrettyDecoded), QUrl::TolerantMode);
     QCOMPARE(url.toString(QUrl::DecodeReserved), theUrl);
     QCOMPARE(url.path(), thePath);
+
+    // local file paths can be fully decoded without loss
+    url.setPath(url.path());
+    QCOMPARE(url.toString(QUrl::DecodeReserved), theUrl);
+    QCOMPARE(url.path(), thePath);
 }
 
 void tst_QUrl::fromLocalFileNormalize_data()
@@ -4150,24 +4160,25 @@ void tst_QUrl::setComponents_data()
                                            << int(Scheme) << "http%61" << Decoded << false
                                            << PrettyDecoded << "" << "";
     QTest::newRow("username-encode") << QUrl("http://example.com")
-                                     << int(UserName) << "h%61llo:world" << Decoded << true
+                                     << int(UserName) << "h%61llo[:]world" << Decoded << true
-                                     << PrettyDecoded << "h%2561llo:world" << "http://h%2561llo%3Aworld@example.com";
+                                     << PrettyDecoded << "h%2561llo[:]world" << "http://h%2561llo%5B%3A%5Dworld@example.com";
     QTest::newRow("password-encode") << QUrl("http://example.com")
-                                     << int(Password) << "h%61llo:world@" << Decoded << true
+                                     << int(Password) << "h%61llo[:]world@" << Decoded << true
-                                     << PrettyDecoded << "h%2561llo:world@" << "http://:h%2561llo:world%40@example.com";
+                                     << PrettyDecoded << "h%2561llo[:]world@" << "http://:h%2561llo%5B:%5Dworld%40@example.com";
     // '%' characters are not permitted in the hostname, these test that it fails to set anything
     QTest::newRow("invalid-host-encode") << QUrl("http://example.com")
                                          << int(Host) << "ex%61mple.com" << Decoded << false
                                          << PrettyDecoded << QString() << QString();
+    // square brackets are force-encoded from decoded forms in the path, query, and fragment
     QTest::newRow("path-encode") << QUrl("http://example.com/foo")
-                                 << int(Path) << "/bar%23" << Decoded << true
+                                 << int(Path) << "/ba[r]%23" << Decoded << true
-                                 << PrettyDecoded << "/bar%2523" << "http://example.com/bar%2523";
+                                 << PrettyDecoded << "/ba%5Br%5D%2523" << "http://example.com/ba%5Br%5D%2523";
     QTest::newRow("query-encode") << QUrl("http://example.com/foo?q")
-                                  << int(Query) << "bar%23" << Decoded << true
+                                  << int(Query) << "ba[r]%23" << Decoded << true
-                                  << PrettyDecoded << "bar%2523" << "http://example.com/foo?bar%2523";
+                                  << PrettyDecoded << "ba%5Br%5D%2523" << "http://example.com/foo?ba%5Br%5D%2523";
     QTest::newRow("fragment-encode") << QUrl("http://example.com/foo#z")
-                                     << int(Fragment) << "bar%23" << Decoded << true
+                                     << int(Fragment) << "ba[r]%23" << Decoded << true
-                                     << PrettyDecoded << "bar%2523" << "http://example.com/foo#bar%2523";
+                                     << PrettyDecoded << "ba%5Br%5D%2523" << "http://example.com/foo#ba%5Br%5D%2523";
     // force decoding
     QTest::newRow("username-decode") << QUrl("http://example.com")
                                      << int(UserName) << "hello%3Aworld%25" << Tolerant << true
@@ -4176,8 +4187,8 @@ void tst_QUrl::setComponents_data()
                                      << int(Password) << "}}>b9o%25kR(" << Tolerant << true
                                      << FullyDecoded << "}}>b9o%kR(" << "http://:%7D%7D%3Eb9o%25kR(@example.com";
     QTest::newRow("path-decode") << QUrl("http://example.com/")
-                                 << int(Path) << "/bar%25foo" << Tolerant << true
+                                 << int(Path) << "/bar%25[foo]" << Tolerant << true
-                                 << FullyDecoded << "/bar%foo" << "http://example.com/bar%25foo";
+                                 << FullyDecoded << "/bar%[foo]" << "http://example.com/bar%25[foo]";
     QTest::newRow("query-decode") << QUrl("http://example.com/foo?qq")
                                   << int(Query) << "bar%25foo" << Tolerant << true
                                   << FullyDecoded << "bar%foo" << "http://example.com/foo?bar%25foo";