[ANN] ParseTree 2.0.0 Released

ruby-talk
1

ParseTree version 2.0.0 has been released!

* <http://www.zenspider.com/ZSS/Products/ParseTree/>

ParseTree is a C extension (using RubyInline) that extracts the parse
tree for an entire class or a specific method and returns it as a
s-expression (aka sexp) using ruby's arrays, strings, symbols, and
integers.

As an example:

   def conditional1(arg1)
     if arg1 == 0 then
       return 1
     end
     return 0
   end

becomes:

   [:defn,
     :conditional1,
     [:scope,
      [:block,
       [:args, :arg1],
       [:if,
        [:call, [:lvar, :arg1], :==, [:array, [:lit, 0]]],
        [:return, [:lit, 1]],
        nil],
       [:return, [:lit, 0]]]]]

* Uses RubyInline, so it just drops in.
* Includes SexpProcessor and CompositeSexpProcessor.
   * Allows you to write very clean filters.
* Includes UnifiedRuby, allowing you to automatically rewrite ruby quirks.
* ParseTree#parse_tree_for_string lets you parse arbitrary strings of ruby.
* Includes parse_tree_show, which lets you quickly snoop code.
   * echo "1+1" | parse_tree_show -f for quick snippet output.
* Includes parse_tree_abc, which lets you get abc metrics on code.
   * abc metrics = numbers of assignments, branches, and calls.
   * whitespace independent metric for method complexity.
* Includes parse_tree_deps, which shows you basic class level dependencies.
* Does not work on the core classes, as they are not ruby (yet).

Changes:

* 2 major enhancements:

   * Rewrite methods completely rewritten. Rewriters:
     * are no longer recursive.
     * are no longer required to empty the sexp coming in.
     * are depth first, so rewriter gets passed everything already normalized.
     * keep rewriting until type doesn't change.
     * are magical goodness.
   * Added UnifiedRuby module to aid others in writing clean SexpProcessors:
     * bmethod/dmethod/fbody unified with defn
     * fcall/vcall unified with call
     * resbody unified with itself (lots of different forms)

* 5 minor enhancements:

   * Add #modules to Module.
   * Add Sexp::for shortcut for Sexp.from_array ParseTree.translate(...).
   * Add parens for :block nodes as appropriate. May be overzealous.
   * Made add_to_parse_tree global for reuse by other C extensions.
   * Made test_ruby2ruby MUCH more rigorous with circular testing.

* 6 bug fixes:

   * Added $TESTING = true to pt_testcase.rb. OOPS!
   * Fixed some circular bugs, mostly by hacking them out, wrt operator precidence.
   * Fixed splat arg processing when arg has no name.
   * Fixed trinary operator.
   * Fixed BMETHOD with no arguments.
   * Removed hacky "self." thing for defs at top level PT use.

* <http://www.zenspider.com/ZSS/Products/ParseTree/>

2

Ryan, thank you for your continuous work on those libraries. I have a
question concerning Ruby's ASTs: do you know a way to get at the parse
tree of the calling method / block and at the node of the call? For
example:

  def look_at_callers_parse_tree
    # what to do here?
  end

  def calling_from_method
    look_at_callers_parse_tree # should get this node...
  end

  def calling_from_block
    yield
  end

  calling_from_block { look_at_callers_parse_tree } # ...and the one
in this block

Regards,
Pit

···

2007/8/1, Ryan Davis <ryand-ruby@zenspider.com>:

ParseTree version 2.0.0 has been released!
(...)

3

May I kindly ask if you could give an example for how to use
parse_tree_for_method? Is there an example available for how to parse a
method, modify the sexpr and feed the result back to the ruby
interpreter, thereby replacing the original method with the patched
version?

Regards,
Thomas.

···

--
Posted via http://www.ruby-forum.com/.

4

not without doing the usual caller() manipulation stuffs, no. Rubinius might be easier in that regard tho.

···

On Aug 1, 2007, at 12:55 , Pit Capitain wrote:

2007/8/1, Ryan Davis <ryand-ruby@zenspider.com>:

ParseTree version 2.0.0 has been released!
(...)

Ryan, thank you for your continuous work on those libraries. I have a
question concerning Ruby's ASTs: do you know a way to get at the parse
tree of the calling method / block and at the node of the call? For
example:

  def look_at_callers_parse_tree
    # what to do here?
  end

  def calling_from_method
    look_at_callers_parse_tree # should get this node...
  end

  def calling_from_block
    yield
  end

  calling_from_block { look_at_callers_parse_tree } # ...and the one
in this block

5

ParseTree is one way. You have to use ruby2ruby + eval to feed the results back in.

···

On Aug 2, 2007, at 04:29 , Tom Link wrote:

May I kindly ask if you could give an example for how to use
parse_tree_for_method? Is there an example available for how to parse a
method, modify the sexpr and feed the result back to the ruby
interpreter, thereby replacing the original method with the patched
version?

6

SuperCaller can do this. It was waiting on a release of PT, because add_to_parse_tree was static.

I need to polish the documentation a bit before I can release.

···

On Aug 1, 2007, at 12:55, Pit Capitain wrote:

2007/8/1, Ryan Davis <ryand-ruby@zenspider.com>:

ParseTree version 2.0.0 has been released!
(...)

Ryan, thank you for your continuous work on those libraries. I have a
question concerning Ruby's ASTs: do you know a way to get at the parse
tree of the calling method / block and at the node of the call? For
example:

  def look_at_callers_parse_tree
    # what to do here?
  end

  def calling_from_method
    look_at_callers_parse_tree # should get this node...
  end

  def calling_from_block
    yield
  end

  calling_from_block { look_at_callers_parse_tree } # ...and the one
in this block

--
Poor workers blame their tools. Good workers build better tools. The
best workers get their tools to do the work for them. -- Syndicate Wars

7

OK. Thanks.

Regards,
Pit

···

2007/8/1, Ryan Davis <ryand-ruby@zenspider.com>:

On Aug 1, 2007, at 12:55 , Pit Capitain wrote:
> (...) do you know a way to get at the parse
> tree of the calling method / block and at the node of the call?

not without doing the usual caller() manipulation stuffs, no.
Rubinius might be easier in that regard tho.

8

Thanks Eric, I'm looking forward to the relase. But maybe I don't need
it anymore, because, while searching the web for SuperCaller, I've
seen on your blog that you seem to work on some of the same ideas as I
do with your inliner experiments :slight_smile:

Regards,
Pit

···

2007/8/4, Eric Hodel <drbrain@segment7.net>:

SuperCaller can do this. It was waiting on a release of PT, because
add_to_parse_tree was static.

I need to polish the documentation a bit before I can release.