From: Luke Graham <spoooq@...>
Date: 2005-02-25T14:21:10+09:00
Subject: Re: increasing ri's searching capibilities and some handy snippets

Its possible to use parsetree and some associated code posted on
http://blog.zenspider.com/archives/ruby/ to display the code for a
function (as sexp, ruby or c, no less).

http://raa.ruby-lang.org/ is a good place to put bits of code.

On Fri, 25 Feb 2005 13:57:59 +0900, Gabriel Horner <cldwalker@chwhat.com> wrote:
> Hi rubyheads,
>   I'm new to the ruby community and am fascinated with its power,simplicity and clean syntax!
> Coming from the perl world where there never seemed to be a standard for documentation subroutines,
> I'm amazed with the ri command for providing such fine-grained searching of classes and their
> methods.  Being a search freak, there are still more ways I'd like to be able to search
> ruby classes,files ,methods and variables. Since I got a lot of questions about rdoc/ri is this
> something I should take to the ruby-core mailing list? If so skip to question 2.
> 
> 1. Extending ri search functions
> 
>         - Display code for a subroutine by extracting it from its file.
>                 I've seen this functionally in html form at http://www.ruby-doc.org/stdlib/
>                 but I don't know how long it'll take to extract the relevant subroutines to do
>                 this from the commandline.
>         - Display documentation for all instance methods of a class at once.
>                 I do this so I can quickly flip and search through to learn how to use a class. Currently
>                 this one-liner works for classes loaded by default. But it would be much handier
>                 if I knew how to map any class to its file (from rdoc generated info) to require the
>                 relevant file.
> 
> ri `ruby -e 'print Object.const_get(ARGV[0].to_s).instance_methods(nil).sort.collect! {|x| ARGV[0]+"."
> +x }.join(" ")' Array`
> 
>         - See modules and classes and their methods defined by filename.
>                 See the sample ruby_file_parser.rb code I've pasted at the end of this email.
> 
>         -Backending ri's searching functions to a database for speed and flexibility in searching.
>                 Any such efforts to import
> 
> 2. If I want to contribute handy snippets of code is the rubyforge.org snippet section the place to
> go?
>  Doesn't seem that it gets much traffic. If I think some snippet is really handy and should come
> distributed with ruby in its sample/ directory who do I mention this to?
> 
> For example, at the end there's a handy snippet, path.rb, I use daily to locate ruby files matching a given regular
> expression.
> 
> Thanks,
> Gabriel
> 
> ##path.rb
> require 'find'
> require 'optparse'
> OPT = {
>         :verbose=>false,
>         :gemdir=>false
> }
> 
> def parseARGV
>         ARGV.options do |opts|
>           script_name = File.basename($0)
>           opts.banner = "Usage: #{script_name} [options] search_term"
> 
>          opts.separator ""
> 
>         opts.on("-g",'--gemdir',"Include gemdir"
>                 ) {|OPT[:gemdir] |}
>         opts.on("-v", "--verbose", "Verbose output","Default: false"
>                   ) { |OPT[:verbose] |}
> 
>           opts.on("-h", "--help",
>                   "Show this help message.") { puts opts; exit }
> 
>           opts.parse!
>         end
> end
> 
> files =[]
> parseARGV
> file = ARGV.shift or raise "No file given"
> GEMDIR = '/usr/local/lib/ruby/gems'
> #or generate with `gem environment gempath`
> searchdir = $:
> searchdir.push(GEMDIR) if OPT[:gemdir]
> searchdir.delete('.')
> i = 0
> 
> Find.find(*searchdir) do |f|
>         puts "#{i}:#{f}" if OPT[:verbose]
>         files.push(f) if f =~ /\/#{file}\.(so|rb)$/
>         i+=1
> end
> puts files.uniq
> __END__
> 
> ##ruby_file_parser.rb
> require 'rdoc/options.rb'
> require 'rdoc/rdoc.rb'
> require 'rdoc/code_objects.rb'
> require 'rdoc/parsers/parse_rb.rb'
> require 'yaml'
> 
> #redefined this subroutine cause the commented out section was throwing a bizarre '*' not defined
> error
> class RubyLex
> class BufferedReader
>         def initialize(content)
>                 if /\t/ =~ content
>                 tab_width = Options.instance.tab_width
>                 #content = content.split(/\n/).map do |line|
>                 #  1 while line.gsub!(/\t+/) { ' ' * (tab_width*$&.length -   $`.length %
> tab_width)}  && $~ #`
>                 #    line
>                 #  end .join("\n")
>                 end
>                 @content   = content
>                 @content << "\n" unless @content[-1,1] == "\n"
>                 @size      = @content.size
>                 @offset    = 0
>                 @hwm       = 0
>                 @line_num  = 1
>                 @read_back_offset = 0
>                 @last_newline = 0
>                 @newline_pending = false
>         end
> end
> end
> 
> $final = {}
> def index_class(cls)
>         class_name = cls.full_name
>         $final[class_name] = cls.method_list.map {|x| x.name } if ! $final[class_name]
>         cls.each_classmodule { |x| index_class(x) }
> end
> 
> file = ARGV.shift or raise "no file given"
> content = File.open(file, "r") {|f| f.read}
> #P = RDoc::RubyParser.new(RDoc::TopLevel.new(file),file,content,Options.instance,RDoc::Stats.new)
> P =
> RDoc::ParserFactory.parser_for(RDoc::TopLevel.new(file),file,content,Options.instance,RDoc::Stats.new)
> P.scan
> rclasses = RDoc::TopLevel.all_classes_and_modules
> classes = rclasses.map {|x| x.full_name }
> 
> for rc in rclasses
>         index_class(rc)
> end
> 
> puts YAML::dump($final)
> __END__
> 
> --
> my looovely website -- http://www.chwhat.com
> BTW, IF chwhat.com goes down email me at gabriel.horner@cern.ch
> 
> 


-- 
spooq