fix: regexp match script tag ignore comment

[btea]

Description

fix #20340

[sapphi-red]

I don't think stripLiteral works with HTML files.
Maybe we can use parse5.Tokenizer to do something similar to stripLiteral but I'm not sure if that is performant than parsing + traversing the HTML directly.

[btea]

You are right. The local test result was correct by accident. The type value in the script tag was removed, resulting in the regular expression not matching.

Can we add another regular expression to match the comment content in HTML? 🤔

[sapphi-red]

I guess we need to at least tokenize it to properly handle comments. In other words, I don't think regex is sufficient.