I attempted to use Tailor to analyse this open-source app. After 10 minutes of maxing out my 4 cores JVM OOMed. I assume this is a limitation based on ANTLR-generated parser being used but I was wondering if you can think of some ways of improving performance by at least an order of magnitude? ;) I'm afraid I don't have much experience to provide meaningful input but one thing that limited memory consumption (to some 1.2GB) when I played around with the grammar was something like this:

private static ParseTree parseTreeFrom(String content) {
  CharStream input = new UnbufferedCharStream(new StringReader(content));
  SwiftLexer lex = new SwiftLexer(input);
  lex.setTokenFactory(new CommonTokenFactory(true));
  TokenStream tokens = new UnbufferedTokenStream<CommonToken>(lex);
  SwiftParser parser = new SwiftParser(tokens);
  parser.setTrimParseTree(true);
  return parser.topLevel();
}

That did not curb insane CPU usage, though. I'm wondering if you have any insights.