Small refactor for template parser

[defcc]

1. Remove the unnecessary argument of parseEndTag.
2. Store the lowerCasedTag in the stack and use it directly.
3. Remove isSpecialTag logic for simplification.

As for isSpecialTag logic, the original implementation checks the tagName and its attribute. If it's a template with lang attribute, it will skip to parse the content inside the template. I think maybe we could simplify it, because the template element is guaranteed by the SFC parser.

const hasLang = attr => attr.name === 'lang' && attr.value !== 'html'

For this line, is it an edge case? There is no proper test case about it, I am not very sure.

Please let me know if anything I missed.

[yyx990803]

Ah cool. For the template tag, the special check was introduced here, but it is no longer necessary because we have improved the HTML parser to forgivingly handle < symbols inside templates now.