When a <script> tag contains something that looks like markup as the first token, the < is not included from element.raw and element.data. This causes problems when parsing templates that use the type='text/html' hack, for example:
... which makes the Text element come out as div></div>.
The above is also included as a test case.
I ran into this issue with jsdom, which still uses htmlparser 1.x.
Fixed parsing of an HTML tag as the first thing inside a <script>.
Whoops, specified the wrong target branch