Fails on too-short UTF-8 sequences consumed from seekable Streams. (#47)

#101 tracks the equivalent fix for non-seekable streams.
amazon-ion · Mar 31, 2020 · 89edf02 · 89edf02
1 parent d01176a
commit 89edf02
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 7 deletions.
diff --git a/Amazon.IonDotnet.Tests/Integration/VectorBad.cs b/Amazon.IonDotnet.Tests/Integration/VectorBad.cs
@@ -29,9 +29,8 @@ public class VectorBad
     {
         private static readonly HashSet<string> Excludes = new HashSet<string>
         {
-            "shortUtf8Sequence_1.ion",
-            "shortUtf8Sequence_2.ion",
-            "shortUtf8Sequence_3.ion"
+            // To exclude a test file of ion-test submodule from running, add the 
+            // test file with its extension here. For example: "test.ion"
         };
 
         private static readonly DirectoryInfo IonTestDir = DirStructure.IonTestDir();

diff --git a/Amazon.IonDotnet/Internals/Text/UnicodeStream.cs b/Amazon.IonDotnet/Internals/Text/UnicodeStream.cs
@@ -1,16 +1,15 @@
 using System;
 using System.Collections.Generic;
-using System.Diagnostics;
 using System.IO;
 using System.Text;
-using Amazon.IonDotnet.Utils;
 
 namespace Amazon.IonDotnet.Internals.Text
 {
     internal class UnicodeStream : TextStream
     {
         private readonly StreamReader _streamReader;
         private readonly Stack<int> _unreadStack;
+        private long? remainingChars = null;
 
         public UnicodeStream(Stream inputStream) : this(inputStream, Encoding.UTF8)
         {
@@ -20,9 +19,12 @@ public UnicodeStream(Stream inputStream, Encoding encoding)
         {
             if (!inputStream.CanRead)
                 throw new ArgumentException("Input stream must be readable", nameof(inputStream));
-
             _streamReader = new StreamReader(inputStream, encoding);
             _unreadStack = new Stack<int>();
+            if (inputStream.CanSeek)
+            {
+                remainingChars = inputStream.Length;
+            }
         }
 
         public UnicodeStream(Stream inputStream, Span<byte> readBytes)
@@ -44,6 +46,7 @@ public UnicodeStream(Stream inputStream, Encoding encoding, Span<byte> readBytes
             _streamReader = new StreamReader(inputStream, encoding);
             if (inputStream.CanSeek)
             {
+                remainingChars = inputStream.Length;
                 InputStream.Seek(-readBytes.Length, SeekOrigin.Current);
                 return;
             }
@@ -57,7 +60,18 @@ public UnicodeStream(Stream inputStream, Encoding encoding, Span<byte> readBytes
 
         public override int Read()
         {
-            return _unreadStack.Count > 0 ? _unreadStack.Pop() : _streamReader.Read();
+            var value = _unreadStack.Count > 0 ? _unreadStack.Pop() : _streamReader.Read();
+
+            if (remainingChars.HasValue)
+            {
+                remainingChars--;
+                if (_streamReader.CurrentEncoding == Encoding.UTF8)
+                {
+                    IsValidUTF8Character();
+                }
+            }
+
+            return value;
         }
 
         public override void Unread(int c)
@@ -72,5 +86,13 @@ public override void Unread(int c)
         }
 
         private Stream InputStream => _streamReader.BaseStream;
+
+        private void IsValidUTF8Character()
+        {
+            if (remainingChars > 0 && _streamReader.Peek() == -1)
+            {
+                throw new IonException("Input stream is not a valid UTF-8 stream.");
+            }
+        }
     }
 }