Unexpected parsing with uppercase DOCTYPE #815
Description
From my blog post detailing the issue. Both of the following two test case documents trigger Info: Document content looks like HTML5 message but produce different results (see below).
Test case 1 (lower case doctype):
echo '<!DOCTYPE html><a href="#"><p>Text</p></a>' | tidy
Test case 1 output (works as expected):
<!DOCTYPE html>
<html>
<body>
<a href="#">
<p>Text</p>
</a>
</body>
</html>
Test case 2 (upper case doctype):
echo '<!DOCTYPE HTML><a href="#"><p>Text</p></a>' | tidy
Test case 2 output (broken; uses legacy pre-html5 parsing):
<!DOCTYPE html>
<html>
<body>
<a href="#"></a>
<p>Text</p>
</body>
</html>
The spec clearly states that the DOCTYPE should be parsed case-insensitively.
(Output in both examples have been trimmed to focus on the differences.)