LALRPOP for parsing AS3?

help
1

I've an AS3 tokenizer written in Rust and have considered using LALRPOP for parsing the AS3 language with a bunch of new compatible features to add.

I've implemented a parser for a similiar language several times using a handwritten approach, similiar to jQuery's Esprima and luaparse. Due to destructuring patterns and arrow functions intersecting in syntax, handwritting is very bug-prone, so I was wishing I could just use a parser generator.

My questions:

XML input goals

Since AS3 inherits the E4X standard (ECMAScript for XML), it has more lexical input goals:

  • InputElementXMLTag
  • InputElementXMLContent

I've different methods for scanning already: scan_ie_div, scan_ie_xml_tag, and scan_ie_xml_content, as well as one for regular expressions to be used ocasionally.

For compliance with the standard, I want the XMLWhitespace nonterminal to be treated as a token.

Will LALRPOP interpret my whitespace correctly?

Diagnostics and line numbers

  1. Should I continue to map line numbers to line offsets inside a Vec<usize>. Does LALRPOP gather the line numbers already when printing diagnostics?
  2. I also built my source locations (Location) such that they additionally contained line numbers. Necessary when using LALRPOP?
  3. I built a Diagnostic structure, which is added to a vector for every respective Source. Is this compatible with LALRPOP?