Ignore the parser translator lexer for sorbet

The whole while switch/case construct makes typechecking difficult because
sorbet does not recognize that outside variables
may change their value in individual switches.

lib/prism/translation/parser/lexer.rb

@@ -238,8 +238,8 @@ def to_a
           index = 0
           length = lexed.length
 
-          heredoc_stack = Array.new
-          quote_stack = Array.new
+          heredoc_stack = []
+          quote_stack = []
 
           # The parser gem emits the newline tokens for comments out of order. This saves
           # that token location to emit at a later time to properly line everything up.

rakelib/typecheck.rake

@@ -37,6 +37,7 @@ namespace :typecheck do
         - ./lib/prism/node_ext.rb
         - ./lib/prism/parse_result.rb
         - ./lib/prism/visitor.rb
+        - ./lib/prism/translation/parser/lexer.rb
         - ./lib/prism/translation/ripper.rb
         - ./lib/prism/translation/ripper/sexp.rb
         - ./lib/prism/translation/ruby_parser.rb