From: Felix von Delius <fvdelius@...>
Date: 2001-04-22T22:40:03+09:00
Subject: [ruby-talk:14019] Re: regexp

I'm interrested in this question, too:

What side effects on the existing syntax would came up it have if a regular 
expression match (using match() or =~) checks for a block given and yields 
that block (in case the expression matches) with the regexp-groups as 
arguments?

This way you could express code like this:

  m = /^(\d+),(\d+)$/.match(line)
  if m then
    first, second = m[1,2]
    print ">>#{first} >>#{second}\n"
  end

much shorter and readable like this:

  /^(\d+),(\d+)$/.match(line) do | first, second |
    print ">>#{first} >>#{second}\n"
  end

Wouldn't it be a large improvement? Would it break existing code?

-Felix

much easied

James Ponder wrote:
> On Sat, Apr 21, 2001 at 10:36:51PM +0900, ts wrote:
>> >>>>> "J" == James Ponder <james@squish.net> writes:
>> J> var =~ /^(.+)=(.+)$/ do |x,y|
>> 
>>   Invalid syntax actually.
> 
> You think it would be too horrible if ruby allowed that syntax and called
> method =~ with the block?  Does that clash with existing syntax rules?
> 
> What do I know.  Just a suggestion :)
> 
> 
> Best wishes, James