Check encoding before parsing; throw exception if not UTF-8

CHANGELOG.md

@@ -4,6 +4,10 @@ Updates should follow the [Keep a CHANGELOG](https://keepachangelog.com/) princi
 
 ## [Unreleased][unreleased]
 
+### Added
+
+ - Added check to ensure Markdown input is valid UTF-8 (#401, #405)
+
 ## [1.2.2] - 2019-01-15
 
 This release contains the same changes as 1.1.3:

src/Converter.php

@@ -50,10 +50,10 @@ public function __construct(DocParserInterface $docParser, ElementRendererInterf
      *
      * @param string $commonMark
      *
-     * @return string
-     *
      * @throws \RuntimeException
      *
+     * @return string
+     *
      * @api
      */
     public function convertToHtml(string $commonMark): string
@@ -70,9 +70,9 @@ public function convertToHtml(string $commonMark): string
      *
      * @param string $commonMark
      *
-     * @return string
-     *
      * @throws \RuntimeException
+     *
+     * @return string
      */
     public function __invoke(string $commonMark): string
     {

src/DocParser.php

@@ -20,6 +20,7 @@
 use League\CommonMark\Block\Element\Paragraph;
 use League\CommonMark\Block\Element\StringContainerInterface;
 use League\CommonMark\Event\DocumentParsedEvent;
+use League\CommonMark\Exception\UnexpectedEncodingException;
 
 final class DocParser implements DocParserInterface
 {
@@ -71,15 +72,16 @@ private function preProcessInput(string $input): array
     /**
      * @param string $input
      *
-     * @return Document
-     *
      * @throws \RuntimeException
+     *
+     * @return Document
      */
     public function parse(string $input): Document
     {
         $document = new Document();
         $context = new Context($document, $this->environment);
 
+        $this->assertValidUTF8($input);
         $lines = $this->preProcessInput($input);
         foreach ($lines as $line) {
             $context->setNextLine($line);
@@ -250,4 +252,11 @@ private function setAndPropagateLastLineBlank(ContextInterface $context, Cursor
             $container = $container->parent();
         }
     }
+
+    private function assertValidUTF8(string $input)
+    {
+        if (!\mb_check_encoding($input, 'UTF-8')) {
+            throw new UnexpectedEncodingException('Unexpected encoding - UTF-8 or ASCII was expected');
+        }
+    }
 }

src/DocParserInterface.php

@@ -18,9 +18,9 @@ interface DocParserInterface
     /**
      * @param string $input
      *
-     * @return Document
-     *
      * @throws \RuntimeException
+     *
+     * @return Document
      */
     public function parse(string $input): Document;
 }

src/Exception/UnexpectedEncodingException.php

@@ -0,0 +1,16 @@
+<?php
+
+/*
+ * This file is part of the league/commonmark package.
+ *
+ * (c) Colin O'Dell <colinodell@gmail.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace League\CommonMark\Exception;
+
+final class UnexpectedEncodingException extends \RuntimeException
+{
+}

tests/unit/CommonMarkConverterTest.php

@@ -60,4 +60,13 @@ public function testEnvironmentAndConfigConstructor()
 
         $this->assertSame($mockEnvironment, $environment);
     }
+
+    /**
+     * @expectedException \League\CommonMark\Exception\UnexpectedEncodingException
+     */
+    public function testConvertingInvalidUTF8()
+    {
+        $converter = new CommonMarkConverter();
+        $converter->convertToHtml("\x09\xca\xca");
+    }
 }

tests/unit/DocParserTest.php

@@ -0,0 +1,30 @@
+<?php
+
+/*
+ * This file is part of the commonmark-php package.
+ *
+ * (c) Colin O'Dell <colinodell@gmail.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace League\CommonMark\Tests\Unit;
+
+use League\CommonMark\DocParser;
+use League\CommonMark\Environment;
+use PHPUnit\Framework\TestCase;
+
+class DocParserTest extends TestCase
+{
+    /**
+     * @expectedException \League\CommonMark\Exception\UnexpectedEncodingException
+     */
+    public function testParsingWithInvalidUTF8()
+    {
+        $environment = Environment::createCommonMarkEnvironment();
+        $docParser = new DocParser($environment);
+
+        $docParser->parse("\x09\xca\xca");
+    }
+}