From: Jason Sweat <jason.sweat@...>
Date: 2006-07-08T10:52:44+09:00
Subject: Re: make this code more Rubyish

On 7/7/06, dblack@wobblini.net <dblack@wobblini.net> wrote:
> I think there's a logic flaw there.  If you have two directories with
> the same diff count, the second one will clobber the hash entry from
> the first one.

Yes, I actually noticed that as I was running it on some other
projects.  Amazing how your first sample can taint your perspective.

Does this iteration look a bit less like line noise?

require 'scanf'

def sloc_scandir(dir)
  /^SLOC.*?$(.*(?=^Totals grouped by))/im.match(
 		%x{sloccount --addlangall #{dir} 2>/dev/null}
  )[1].to_s
end

def get_counts(lines)
  res = {}
  lines.scanf("%i%s%s") do |count, thing|
    next unless count > 0
    res[thing] = count
  end
  res
end

class SlocDir
	attr_accessor :name, :dev, :prd
	def initialize(name, dev)
	  @name = name
	  @dev = dev
	end
	def diff
	  @dev - @prd
	end
	def render
	  "#{@name}\t#{@prd}\t#{@dev}\t#{diff}"
	end
end

dirs = {}
get_counts(sloc_scandir('.')).each_pair { |k,v| dirs[k] = SlocDir.new(k,v) }
get_counts(sloc_scandir(`cat .prod_dir`)).each_pair { |k,v| dirs[k].prd = v }

puts "Dir\tProd\tDev\tDiff"
dirs.values.sort{ |x,y| y.diff <=> x.diff }.each { |d| puts d.render }



Regards,
Jason
http://blog.casey-sweat.us/