Lexer: Efficient error handling

[ruiconti]

As of now, there is little to no error handling in lexer. This needs to be done for all kinds of tokens:

• Numeric literals
• String literals
• Punctuators
• Template literals
• RegExp literals
• Identifiers

Also, it is the right moment to add more meaningful messages and precise source position report.

https://262.ecma-international.org/13.0/#sec-error-handling-and-language-extensions