ruby_parser version 3.2.0 has been released!
Published 2013-07-02 @ 20:55
ruby_parser (RP) is a ruby parser written in pure ruby (utilizing racc–which does by default use a C extension). RP’s output is the same as ParseTree’s output: s-expressions using ruby’s arrays and base types.
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)))
Changes:
3.2.0 / 2013-07-02
-
1 major enhancement:
- Added (rough draft) 2.0 support. Still missing some small / rare things.
-
12 minor enhancements:
- Added %i(symbol-names…) support. (%I too)
- Added 140 more tests, jumping test count from 1376 to 2143. Yay for test reuse!
- Added RubyLexer#brace_nest.
- Added compare20 rake task to diff the grammar architecture against MRI.
- Added lpar_beg and paren_nest to lexer to track state of parens in stabbies
- Added shadow nodes for scoped block args.
- Compound RubyParser now defaults to 2.0.
- Fixed rake to < 10, because 10’s file dependency handling is so very broken.
-
Made it possible to specify version in bin/ruby_parse_extract_error w/ -v 18 19 20 - Refactored to RubyParserStuff::ENCODING_ORDER to allow custom tweaking of encoding guessing. (samlown)
- Switched
rake debug
to default to 2.0. - Translated some fixes across 1.8 and 1.9 from 2.0.
-
42 bug fixes:
- 2.0: Fixed a number of block args scenarios w/ kwargs
- 2.0: Fixed args_tail mismatching against lexer.
- 2.0: Fixed assocs to return a hash node.
- 2.0: Fixed f_block_kw production.
- 2.0: Fixed f_block_kwarg production.
- 2.0: Fixed handling of stabby proc args in parens.
- 2.0: Fixed lexing of kwsplat nodes.
- 2.0: Implemented kwsplat nodes.
- Added tUBANG to lexer.
- Apparently ruby doesn’t warn for escaped octal that goes above 7. wtf.
- Cleaned up a LOT of arg handling (block and defn, not calls) by using #args.
- ESC_RE is set to unicode. This seems problematic. shrug
- Either found a bug in MRI and/or fixed paren_nest count for []=? methods.
- Extended IDENT_CHAR_RE on 1.9+ to top out at \u{10ffff}… because we NEED a million unicode chars.
- Fixed % strings with interpolation.
- Fixed BEGIN {} to return a sexp.
- Fixed a[] += b. (stormbrew)
- Fixed arg_blk_pass to allow for sub-args nodes.
- Fixed assignable to allow for sexps to be passed to it.
- Fixed assoc args in 1.9.
- Fixed block_command and block_call (eg a.b c d) to #to_sym their args properly.
- Fixed bug in compound RubyParser so it rescues RubyParser::SyntaxError.
- Fixed env registration of cdecls.
- Fixed lex value for { when expr_endfn.
- Fixed lex_state for close paren/brace/bracket.
- Fixed lex_state transition for 1.9 if we lexed a defn name. Only 1.8 is odd.
- Fixed lexer problem with state mgmt of identifiers that also have registered var name.
- Fixed lexing of “1 *\n” to have the correct lex_state.
- Fixed lexing of heredocs vs chevron for some lex_states.
- Fixed op_asgn nodes to #to_sym their args properly.
- Fixed optional value block args.
- Fixed parsing of ENCODING on ruby 1.8 (vcall).
- Fixed some oddity where 1.9 lexing was blowing up on “0o”. Seems invalid now.
- Fixed strings with escaped octals > 128. Also… wtf.
- Fixed support for empty symbol (wtf?).
- Lexer is now declared UTF-8 internally. Hopefully this will fix the encoding mess.
- Made UTF_8 the default guess on encodings when it isn’t explicit.
- Parsing of ENCODING on ruby 1.9+ (in ruby 1.9+) is now colon2 sexp. (whitequark)
- Renamed RubyLexer#nest to string_nest
- RubyLexer#unescape ignores bad octal/hex and returns unicode strings.
- Switched a number of lexical constructs to use IDENT_CHAR_RE instead of \w. I wish there were something cleaner for regexps + unicode.
- Switched ruby_parse_extract_error to use binread.
- rdoc: http://docs.seattlerb.org/ruby_parser
- home: https://github.com/seattlerb/ruby_parser
- bugs: https://github.com/seattlerb/ruby_parser/issues