From 35d3f36d2d8661fa17eb3a374ee1f4c9d868d08a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?S=C3=A9bastien=20Ros?= <sebastienros@gmail.com>
Date: Tue, 7 Jan 2025 11:51:53 -0800
Subject: [PATCH] Fix escaped numbers parsing (#192)

---
 .github/workflows/build.yml       |  3 +++
 .github/workflows/publish.yml     |  3 +++
 src/Parlot/Scanner.cs             | 18 +++++++++++------
 test/Parlot.Tests/ScannerTests.cs | 33 ++++++++++++++++++++++++++++++-
 4 files changed, 50 insertions(+), 7 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 43234a9..8a7ee02 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -25,6 +25,9 @@ jobs:
     - uses: actions/checkout@v4
     - uses: actions/setup-dotnet@v4
       with:
+        dotnet-version: | 
+          6.0.x
+          8.0.x
         global-json-file: global.json
 
     - name: Test
diff --git a/.github/workflows/publish.yml b/.github/workflows/publish.yml
index f0ec025..fee65f1 100644
--- a/.github/workflows/publish.yml
+++ b/.github/workflows/publish.yml
@@ -17,6 +17,9 @@ jobs:
     - uses: actions/checkout@v4
     - uses: actions/setup-dotnet@v4
       with:
+        dotnet-version: | 
+          6.0.x
+          8.0.x
         global-json-file: global.json
     
     - name: Test
diff --git a/src/Parlot/Scanner.cs b/src/Parlot/Scanner.cs
index 68c76a6..273c415 100644
--- a/src/Parlot/Scanner.cs
+++ b/src/Parlot/Scanner.cs
@@ -556,6 +556,8 @@ private bool ReadQuotedString(char quoteChar, out ReadOnlySpan<char> result)
             // We can read Eof if there is an escaped quote sequence and no actual end quote, e.g. "'abc\'def"
             if (Cursor.Eof)
             {
+                Cursor.ResetPosition(start);
+
                 result = [];
                 return false;
             }
@@ -585,11 +587,13 @@ private bool ReadQuotedString(char quoteChar, out ReadOnlySpan<char> result)
                         // https://stackoverflow.com/a/32175520/142772
                         // exactly 4 digits
 #if NET8_0_OR_GREATER
-                        var lastHexIndex = Cursor.Span.Slice(0, 4).LastIndexOfAny(Character._hexDigits);
-                        var isValidUnicode = lastHexIndex == 3;
+                        var allHexDigits = Cursor.Span.Length > 4 && Cursor.Span.Slice(1, 4).IndexOfAnyExcept(Character._hexDigits) == -1;
+                        var isValidUnicode = allHexDigits;
 
                         if (!isValidUnicode)
                         {
+                            Cursor.ResetPosition(start);
+
                             result = [];
                             return false;
                         }
@@ -629,19 +633,21 @@ private bool ReadQuotedString(char quoteChar, out ReadOnlySpan<char> result)
                         break;
                     case 'x':
 
-                        // At least two digits
+                        // At least one digits
 #if NET8_0_OR_GREATER
-                        lastHexIndex = Cursor.Span.Slice(0, 4).LastIndexOfAny(Character._hexDigits);
-                        var isValidHex = lastHexIndex > 0;
+                        var firstNonHexDigit = Cursor.Span.Length > 1 ? Cursor.Span.Slice(1).IndexOfAnyExcept(Character._hexDigits) : -1;
+                        var isValidHex = firstNonHexDigit > 0;
 
                         if (!isValidHex)
                         {
+                            Cursor.ResetPosition(start);
+
                             result = [];
                             return false;
                         }
 
                         // Advance the cursor for the read digits
-                        Cursor.Advance(lastHexIndex + 1);
+                        Cursor.Advance(firstNonHexDigit);
 #else
                         var isValidHex = false;
 
diff --git a/test/Parlot.Tests/ScannerTests.cs b/test/Parlot.Tests/ScannerTests.cs
index 1af848b..c215346 100644
--- a/test/Parlot.Tests/ScannerTests.cs
+++ b/test/Parlot.Tests/ScannerTests.cs
@@ -277,10 +277,41 @@ public void ShouldReadStringsWithLineBreaks(string text, string expected)
     [InlineData("' \\xa0 ' ", "' \\xa0 '")]
     [InlineData("' \\xfh ' ", "' \\xfh '")]
     [InlineData("' \\u1234 ' ", "' \\u1234 '")]
-
     public void ShouldReadUnicodeSequence(string text, string expected)
     {
         new Scanner(text).ReadQuotedString(out var result);
         Assert.Equal(expected, result);
     }
+
+    [Theory]
+    [InlineData("'\\u'")]
+    [InlineData("'\\u1'")]
+    [InlineData("'\\u12'")]
+    [InlineData("'\\u123'")]
+    [InlineData("'\\ug'")]
+    [InlineData("'\\u1g'")]
+    [InlineData("'\\u12g'")]
+    [InlineData("'\\u123g'")]
+    [InlineData("'\\x'")]
+    [InlineData("'\\xg'")]
+    public void ShouldNotParseInvalidEscapedNumbers(string input)
+    {
+        var s = new Scanner(input);
+        Assert.False(s.ReadQuotedString());
+        Assert.Equal(0, s.Cursor.Position.Offset);
+    }
+
+    [Theory]
+    [InlineData("'\\u1234'")]
+    [InlineData("'\\u12345'")]
+    [InlineData("'\\x1'")]
+    [InlineData("'\\x12'")]
+    [InlineData("'\\x123'")]
+    [InlineData("'\\x1234'")]
+    [InlineData("'\\x1234g'")]
+    public void ShouldParseValidEscapedNumbers(string input)
+    {
+        var s = new Scanner(input);
+        Assert.True(s.ReadQuotedString());
+    }
 }