From: Chris <chris@...>
Date: 2002-04-16T05:42:46+09:00
Subject: Re: need help on using Ruby to replace some SED expressions

In article <90250227.0204151017.1503fb89@posting.google.com>,
 kackson@yahoo.com (kackson) wrote:

> I tried sed. But it is getting really tough when I keep realizing I've
> things to expand and expand and expand into a huge chunk of regular
> expressions that now become completely unreadable and hard to
> maintain.
> So, as a beginner of Ruby, any advise on the question below is
> apprecaited.
> 
> I need to:
> 1. read an input file
> 2. For everyline, if the FIRST word has a dot '.', replace it with $.
> Note, the first word may/maynot have been indented.
> 3. For the rest of the line, replace all occurance of '@' with '^'.
> 4. If there is something like '(2 + 3)' somewhere along the line,
> evaluate it and replace the entire expression i.e. (2 + 3) with the
> result, i.e. 5. Note, the brackets are gone. And it could be +, *, -,
> or divide. If it is even possible for mathmatical routines such as
> log/sin, etc. That would be even better.

This seems like it should work, but hangs when I try to run it...

#!/usr/local/bin/ruby

class String
   def eval_math!
      while self =~ /(log|sin|cos|tan)\s+(\d+)/
         sub! $&, eval("Math.#{$1} #{$2}").to_s
      end
      re = /\(\s*[\-+]?\d+(\.\d+)?(\s*[+-\/*]\s*[\-+]?\d+(\.\d+)?)*\s*\)/
      while self =~ re
         x = $&
         sub! x, eval(x).to_s
      end
      self
   end
end

ARGV.each do |filename|
   file = open(filename, "r")
   text = file.readlines
   file.close
   text.collect! do |line|
      line.sub!(/^(\s*\S*)\.(\S*)/, "#{$1}$#{$2}")
      line.sub!("@", "^")
      line.eval_math!
   end
   file = open(filename, "w")
   file.write(text.join)
   file.close
end