From 3746eb8412ea42d7e3c519926460482530782a75 Mon Sep 17 00:00:00 2001
From: Samuel Gaist <samuel.gaist@idiap.ch>
Date: Wed, 1 Aug 2018 00:28:26 +0200
Subject: QRegularExpression: refactor wildcard translation

This patch refactors the wildcardToRegularExpression method to generate
a simpler regular expression. It also fixes some shortcomings of the
previous implementation.

Tests have been updated to ensure all cases are properly supported.

Change-Id: I454e3fe8fe0bb663b2f319d6fa2fa8aec626c50d
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
---
 src/corelib/tools/qregularexpression.cpp           | 59 +++++++++++++---------
 .../qregularexpression/tst_qregularexpression.cpp  | 19 +++++++
 2 files changed, 54 insertions(+), 24 deletions(-)

diff --git a/src/corelib/tools/qregularexpression.cpp b/src/corelib/tools/qregularexpression.cpp
index 17caaabc97..1bd06a73cd 100644
--- a/src/corelib/tools/qregularexpression.cpp
+++ b/src/corelib/tools/qregularexpression.cpp
@@ -1874,6 +1874,9 @@ QString QRegularExpression::escape(const QString &str)
     \since 5.12
 
     Returns a regular expression representation of the given glob \a pattern.
+    The transformation is targeting file path globbing, which means in particular
+    that path separators receive special treatment. This implies that it is not
+    just a basic translation from "*" to ".*".
 
     \snippet code/src_corelib_tools_qregularexpression.cpp 31
 
@@ -1917,19 +1920,35 @@ QString QRegularExpression::wildcardToRegularExpression(const QString &pattern)
 {
     const int wclen = pattern.length();
     QString rx;
+    rx.reserve(wclen + wclen / 16);
     int i = 0;
-    bool hasNegativeBracket = false;
     const QChar *wc = pattern.unicode();
 
+#ifdef Q_OS_WIN
+    const QLatin1Char nativePathSeparator('\\');
+    const QLatin1String starEscape("[^/\\\\]*");
+    const QLatin1String questionMarkEscape("[^/\\\\]");
+#else
+    const QLatin1Char nativePathSeparator('/');
+    const QLatin1String starEscape("[^/]*");
+    const QLatin1String questionMarkEscape("[^/]");
+#endif
+
     while (i < wclen) {
         const QChar c = wc[i++];
         switch (c.unicode()) {
         case '*':
-            rx += QLatin1String(".*");
+            rx += starEscape;
             break;
         case '?':
-            rx += QLatin1Char('.');
+            rx += questionMarkEscape;
+            break;
+        case '\\':
+#ifdef Q_OS_WIN
+        case '/':
+            rx += QLatin1String("[/\\\\]");
             break;
+#endif
         case '$':
         case '(':
         case ')':
@@ -1943,35 +1962,27 @@ QString QRegularExpression::wildcardToRegularExpression(const QString &pattern)
             rx += c;
             break;
         case '[':
+            rx += c;
             // Support for the [!abc] or [!a-c] syntax
-            // Implements a negative look-behind for one char.
             if (i < wclen) {
-                if (wc[i] == QLatin1Char(']')) {
-                    rx += c;
-                    rx += wc[i++];
-                } else if (wc[i] == QLatin1Char('!')) {
-                    rx += QLatin1String(".(?<");
-                    rx += wc[i++];
-                    rx += c;
-                    hasNegativeBracket = true;
-                } else {
-                    rx += c;
+                if (wc[i] == QLatin1Char('!')) {
+                    rx += QLatin1Char('^');
+                    ++i;
                 }
+
+                if (i < wclen && wc[i] == QLatin1Char(']'))
+                    rx += wc[i++];
+
                 while (i < wclen && wc[i] != QLatin1Char(']')) {
+                    // The '/' appearing in a character class invalidates the
+                    // regular expression parsing. It also concerns '\\' on
+                    // Windows OS types.
+                    if (wc[i] == QLatin1Char('/') || wc[i] == nativePathSeparator)
+                        return rx;
                     if (wc[i] == QLatin1Char('\\'))
                         rx += QLatin1Char('\\');
                     rx += wc[i++];
                 }
-            } else {
-                rx += c;
-            }
-            break;
-        case ']':
-            rx += c;
-            // Closes the negative look-behind expression.
-            if (hasNegativeBracket) {
-                rx += QLatin1Char(')');
-                hasNegativeBracket = false;
             }
             break;
         default:
diff --git a/tests/auto/corelib/tools/qregularexpression/tst_qregularexpression.cpp b/tests/auto/corelib/tools/qregularexpression/tst_qregularexpression.cpp
index 987ca519ee..f520e9742a 100644
--- a/tests/auto/corelib/tools/qregularexpression/tst_qregularexpression.cpp
+++ b/tests/auto/corelib/tools/qregularexpression/tst_qregularexpression.cpp
@@ -2173,6 +2173,7 @@ void tst_QRegularExpression::wildcard_data()
     addRow("?m", "test.html", 6);
     addRow("[*]", "test.html", -1);
     addRow("[?]","test.html", -1);
+    addRow("[?]","test.h?ml", 6);
     addRow("[[]","test.h[ml", 6);
     addRow("[]]","test.h]ml", 6);
     addRow(".h[a-z]ml", "test.html", 4);
@@ -2187,6 +2188,24 @@ void tst_QRegularExpression::wildcard_data()
     addRow(".h[][!]", "test.h]ml", 4);
     addRow(".h[][!]", "test.h[ml", 4);
     addRow(".h[][!]", "test.h!ml", 4);
+
+    addRow("foo/*/bar", "Qt/foo/baz/bar", 3);
+    addRow("foo/(*)/bar", "Qt/foo/baz/bar", -1);
+    addRow("foo/(*)/bar", "Qt/foo/(baz)/bar", 3);
+    addRow("foo/?/bar", "Qt/foo/Q/bar", 3);
+    addRow("foo/?/bar", "Qt/foo/Qt/bar", -1);
+    addRow("foo/(?)/bar", "Qt/foo/Q/bar", -1);
+    addRow("foo/(?)/bar", "Qt/foo/(Q)/bar", 3);
+
+#ifdef Q_OS_WIN
+    addRow("foo\\*\\bar", "Qt\\foo\\baz\\bar", 3);
+    addRow("foo\\(*)\\bar", "Qt\\foo\\baz\\bar", -1);
+    addRow("foo\\(*)\\bar", "Qt\\foo\\(baz)\\bar", 3);
+    addRow("foo\\?\\bar", "Qt\\foo\\Q\\bar", 3);
+    addRow("foo\\?\\bar", "Qt\\foo\\Qt\\bar", -1);
+    addRow("foo\\(?)\\bar", "Qt\\foo\\Q\\bar", -1);
+    addRow("foo\\(?)\\bar", "Qt\\foo\\(Q)\\bar", 3);
+#endif
 }
 
 void tst_QRegularExpression::wildcard()
-- 
cgit v1.2.3