🕷 software releases

by ryan davis



sitemap

ruby_parser version 3.18.0 has been released!

Published 2021-10-27 @ 17:06

ruby_parser (RP) is a ruby parser written in pure ruby (utilizing racc–which does by default use a C extension). It outputs s-expressions which can be manipulated and converted back to ruby via the ruby2ruby gem.

As an example:

def conditional1 arg1
  return 1 if arg1 == 0
  return 0
end

becomes:

s(:defn, :conditional1, s(:args, :arg1),
  s(:if,
    s(:call, s(:lvar, :arg1), :==, s(:lit, 0)),
    s(:return, s(:lit, 1)),
    nil),
  s(:return, s(:lit, 0)))

Tested against 801,039 files from the latest of all rubygems (as of 2013-05):

  • 1.8 parser is at 99.9739% accuracy, 3.651 sigma
  • 1.9 parser is at 99.9940% accuracy, 4.013 sigma
  • 2.0 parser is at 99.9939% accuracy, 4.008 sigma
  • 2.6 parser is at 99.9972% accuracy, 4.191 sigma

Changes:

3.18.0 / 2021-10-27

Holy crap… 58 commits! 2.7 and 3.0 are feature complete. Strings & heredocs have been rewritten.

  • 9 major enhancements:

    • !!! Rewrote lexer (and friends) for strings, heredocs, and %*[] constructs.
    • Massive overhaul on line numbers.
    • Freeze input! Finally!!! No more modifying the input string for heredocs.
    • Overhauled RPStringScanner. Removed OLD compatibility methods!
    • Removed Sexp methods: value, to_sym, add, add_all, node_type, values.
      • value moved to sexp_processor.
    • Removed String#grep monkey-patch.
    • Removed String#lineno monkey-patch.
    • Removed string_to_pos, charpos, etc hacks for ancient ruby versions.
    • Removed unread_many… NO! NO EDITING THE INPUT STRING!
  • 31 minor enhancements:

    • 2.7/3.0: many more pattern edge cases
    • 2.7: Added mlhs = rhs rescue expr
    • 2.7: refactored destructured args (|(k,v)|) and unfactored(?!) case_body/args.
    • 3.0: excessed_comma
    • 3.0: finished most everything: endless methods, patterns, etc.
    • 3.0: refactored / added new pattern changes
    • Added RubyLexer#in_heredoc? (ie, is there old_ss ?)
    • Added RubyLexer#old_ss and old_lineno and removed much of SSStack(ish).
    • Added Symbol#end_with? when necessary
    • Added TALLY and DEBUG options for ss.getch and ss.scan
    • Added ignore_body_comments to make parser productions more clear.
    • Added support for no_kwarg (eg def f(**nil)).
    • Added support for no_kwarg in blocks (eg f { |**nil| }).
    • Augmented generated parser files to have frozen_string_literal comments and fixed tests.
    • Broke out 3.0 parser into its own to ease development.
    • Bumped dependencies on sexp_processor and oedipus_lex.
    • Clean generated 3.x files.
    • Extracted all string scanner methods to their own module.
    • Fixed some precedence decls.
    • Implemented most of pattern matching for 2.7+.
    • Improve lex_state= to report location in verbose debug mode.
    • Made it easier to debug with a particular version of ruby via rake.
    • Make sure ripper uses the same version of ruby we specified.
    • Moved all string/heredoc/etc code to ruby_lexer_strings.rb
    • Remove warning from newer bisons.
    • Sprinkled in some frozen_string_literal, but mostly helped by oedipus bump.
    • Switch to comparing against ruby binary since ripper is buggy.
    • bugs task should try both bug.rb and bad.rb.
    • endless methods
    • f_any_kwrest refactoring.
    • refactored defn/defs
  • 15 bug fixes:

    • Cleaned a bunch of old hacks. Initializing RubyLexer w/ Parser is cleaner now.
    • Corrected some lex_state errors in process_token_keyword.
    • Fixed ancient ruby2 change (use #lines) in ruby_parse_extract_error.
    • Fixed bug where else without rescue only raises on 2.6+
    • Fixed caller for getch and scan when DEBUG=1
    • Fixed comments in the middle of message cascades.
    • Fixed differences w/ symbol productions against ruby 2.7.
    • Fixed dsym to use string_contents production.
    • Fixed error in bdot2/3 in some edge cases. Fixed p_alt line.
    • Fixed heredoc dedenting in the presence of empty lines. (mvz)
    • Fixed some leading whitespace / comment processing
    • Fixed up how class/module/defn/defs comments were collected.
    • Overhauled ripper.rb to deal with buggy ripper w/ yydebug.
    • Removed dsym from literal.
    • Removed tUBANG lexeme but kept it distinct as a method name (eg: def !@).
  • home: https://github.com/seattlerb/ruby_parser
  • bugs: https://github.com/seattlerb/ruby_parser/issues
  • rdoc: http://docs.seattlerb.org/ruby_parser