Fix length check for Regex BOL FindFirstChar optimization (#55574)

For a beginning-of-line anchor, in FindFirstChar we use IndexOf to quickly skip ahead to the next \n. But we neglected to check to see whether that brought us past an explicitly specified end position. This just adds the missing check.
dotnet · Jul 13, 2021 · ce09822 · ce09822
1 parent a38fc2f
commit ce09822
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 12 deletions.
diff --git a/...raries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCompiler.cs b/...raries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCompiler.cs
@@ -1240,19 +1240,21 @@ protected void GenerateFindFirstChar()
                             {
                                 Stloc(newlinePos);
 
-                                // if (newlinePos == -1)
+                                // if (newlinePos == -1 || newlinePos + 1 > runtextend)
                                 // {
                                 //     runtextpos = runtextend;
                                 //     return false;
                                 // }
-                                Label foundNextLine = DefineLabel();
                                 Ldloc(newlinePos);
                                 Ldc(-1);
-                                Bne(foundNextLine);
-                                BrFar(returnFalse);
+                                Beq(returnFalse);
+                                Ldloc(newlinePos);
+                                Ldc(1);
+                                Add();
+                                Ldloc(_runtextendLocal);
+                                Bgt(returnFalse);
 
                                 // runtextpos = newlinePos + 1;
-                                MarkLabel(foundNextLine);
                                 Ldloc(newlinePos);
                                 Ldc(1);
                                 Add();

diff --git a/...ies/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexInterpreter.cs b/...ies/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexInterpreter.cs
@@ -406,7 +406,7 @@ protected override bool FindFirstChar()
                 if (runtextpos > runtextbeg && runtext![runtextpos - 1] != '\n')
                 {
                     int newline = runtext.IndexOf('\n', runtextpos);
-                    if (newline == -1)
+                    if (newline == -1 || newline + 1 > runtextend)
                     {
                         runtextpos = runtextend;
                         return false;
@@ -457,7 +457,7 @@ protected override bool FindFirstChar()
                     if (!_code.LeadingCharClasses[0].CaseInsensitive)
                     {
                         // singleton, left-to-right, case-sensitive
-                        int i = runtext.AsSpan(runtextpos, runtextend - runtextpos).IndexOf(ch);
+                        int i = span.IndexOf(ch);
                         if (i >= 0)
                         {
                             runtextpos += i;

diff --git a/src/libraries/System.Text.RegularExpressions/tests/Regex.Match.Tests.cs b/src/libraries/System.Text.RegularExpressions/tests/Regex.Match.Tests.cs
@@ -152,6 +152,9 @@ public static IEnumerable<object[]> Match_Basic_TestData()
             // Using beginning/end of string chars \A, \Z: Actual - "\\Aaaa\\w+zzz\\Z"
             yield return new object[] { @"\Aaaa\w+zzz\Z", "aaaasdfajsdlfjzzza", RegexOptions.None, 0, 18, false, string.Empty };
 
+            // Anchors and multiline
+            yield return new object[] { @"^A$", "ABC\n", RegexOptions.Multiline, 0, 2, false, string.Empty };
+
             // Using beginning/end of string chars \A, \Z: Actual - "\\Aaaa\\w+zzz\\Z"
             yield return new object[] { @"\A(line2\n)line3\Z", "line2\nline3\n", RegexOptions.Multiline, 0, 12, true, "line2\nline3" };
 
@@ -813,7 +816,7 @@ public static IEnumerable<object[]> Match_Advanced_TestData()
                 }
             };
 
-            // Mutliline
+            // Multiline
             yield return new object[]
             {
                 "(line2$\n)line3", "line1\nline2\nline3\n\nline4", RegexOptions.Multiline, 0, 24,
@@ -824,7 +827,7 @@ public static IEnumerable<object[]> Match_Advanced_TestData()
                 }
             };
 
-            // Mutliline
+            // Multiline
             yield return new object[]
             {
                 "(line2\n^)line3", "line1\nline2\nline3\n\nline4", RegexOptions.Multiline, 0, 24,
@@ -835,7 +838,7 @@ public static IEnumerable<object[]> Match_Advanced_TestData()
                 }
             };
 
-            // Mutliline
+            // Multiline
             yield return new object[]
             {
                 "(line3\n$\n)line4", "line1\nline2\nline3\n\nline4", RegexOptions.Multiline, 0, 24,
@@ -846,7 +849,7 @@ public static IEnumerable<object[]> Match_Advanced_TestData()
                 }
             };
 
-            // Mutliline
+            // Multiline
             yield return new object[]
             {
                 "(line3\n^\n)line4", "line1\nline2\nline3\n\nline4", RegexOptions.Multiline, 0, 24,
@@ -857,7 +860,7 @@ public static IEnumerable<object[]> Match_Advanced_TestData()
                 }
             };
 
-            // Mutliline
+            // Multiline
             yield return new object[]
             {
                 "(line2$\n^)line3", "line1\nline2\nline3\n\nline4", RegexOptions.Multiline, 0, 24,