From: Michael Neumann <neumann@...>
Date: 2001-07-19T18:36:13+09:00
Subject: [ruby-talk:18107] Re: Best way to prevent infinite loops...

Sean Chittenden wrote:
> > > 	Howdy.  What's the best way to prevent infinite loops in Ruby?  
> > 
> > Don't write one.  :)
> 
> 	I wish I could...  I'm using a recursive regexp to remove all
> <?tmpl.* ?> tags from a template and tags can include tags.
> 
> > I guess I'd ask: what's the driver of the loop? Is it the test for
> > some_action, or is it the count, or both. If the count is incidental,
> > then I'd tuck it away inside the loop (although possibly making it
> > slightly shorter)
> 
> 	In a nut shell, here's what I'm doing:
> 
>     tmpl = large_string
>     tag_re = Regexp.new('\<\?tmpl\.([^\s\?]*)\s*(.*?)\s*\?\>')
>     loop_count = 0
>     while md = tag_re.match(tmpl)
>       if loop_count > @max_num_iteration
>         raise(RuntimeError, "Too many iterations")
>       else
>         loop_count += 1
>       end
>       ...
>       tag = md.to_a[0]
>       ...
>       tmpl.gsub!(/#{Regexp.escape(tag)}/, val)
>       ...
>     end


def loop_max(max_iter)
  i = 0
  catch(:exit) {
    loop {
      if i > max_iter
        raise RuntimeError, "Too many interations"
      else
        yield
      end
      i += 1
    }
  }
end


tmpl = large_string
tag_re = Regexp.new('\<\?tmpl\.([^\s\?]*)\s*(.*?)\s*\?\>')
loop_max(@max_num_iteration) do 
  throw :exit unless md = tag_re.match(tmpl)
  ...
  tag = md.to_a[0]
  ...
  tmpl.gsub!(/#{Regexp.escape(tag)}/, val)
  ...
end

Or instead of "throw" you could use the return value of the block
as exit condition.

Another solution would be to use timeout:

require "timeout"

timeout(10) {
  ...
}


Regards,

  Michael


-- 
Michael Neumann
merlin.zwo InfoDesign GmbH
http://www.merlin-zwo.de