HTML reader: don't parse raw HTML inside <code> tag.

Previously '<code><a>x</a></code>' would be parsed as
Code "<a>x</a>", which is not what you want.

src/Text/Pandoc/Readers/HTML.hs

@@ -686,8 +686,8 @@ inline = choice [ charRef
 
 code :: GenParser Char ParserState Inline
 code = try $ do 
-  result <- (htmlOpenTag "code" >> manyTill anyChar (htmlEndTag "code"))
-        <|> (htmlOpenTag "tt"   >> manyTill anyChar (htmlEndTag "tt"))
+  result <- (htmlOpenTag "code" >> manyTill (noneOf "<>") (htmlEndTag "code"))
+        <|> (htmlOpenTag "tt"   >> manyTill (noneOf "<>") (htmlEndTag "tt"))
   -- remove internal line breaks, leading and trailing space,
   -- and decode character references
   return $ Code $ decodeCharacterReferences $ removeLeadingTrailingSpace $ 

tests/html-reader.html

@@ -340,7 +340,7 @@ foo<p>This should just be an HTML comment:</p>
 <p>So is <strong><em>this</em></strong> word.</p>
 <p><strong><em>This is strong and em.</em></strong></p>
 <p>So is <strong><em>this</em></strong> word.</p>
-<p>This is code: <code>></code>, <code>$</code>, <code>\</code>, <code>\$</code>, <code>&lt;html></code>.</p>
+<p>This is code: <code>&gt;</code>, <code>$</code>, <code>\</code>, <code>\$</code>, <code>&lt;html&gt;</code>.</p>
 <hr />
 <h1>Smart quotes, ellipses, dashes</h1>
 <p>"Hello," said the spider. "'Shelob' is my name."</p>
@@ -442,8 +442,8 @@ Email link (nobody [at] nowhere.net)<p><a href="">Empty</a>.</p>
 An e-mail address: nobody [at] nowhere.net<blockquote>
 <p>Blockquoted: <a href="http://example.com/">http://example.com/</a></p>
 </blockquote>
-<p>Auto-links should not occur here: <code>&lt;http://example.com/></code></p>
-<pre><code>or here: &lt;http://example.com/>
+<p>Auto-links should not occur here: <code>&lt;http://example.com/&gt;</code></p>
+<pre><code>or here: &lt;http://example.com/&gt;
 </code></pre>
 <hr />
 <h1>Images</h1>