From: David Flanagan <david@...>
Date: 2007-09-11T14:05:29+09:00
Subject: Re: integer range literals in conditionals with -e option

Nobuyoshi Nakada wrote:
>> Anyone know what the history and current status of this is?  Is the code 
>> in parse.y vestigal, or is  a regresson that this feature stopped 
>> working in 1.8?
> 
> 
> # I'm surprised that nobody noticed it over 2 years.
> 
> I thought of these alternatives:
> 
> a) delay the transformation of NODE_DOT till conditional
>    expressions take place (the following patch)
> 
> b) use Range#=== (as you thought), but this would alter the
>    semantics and make some inconsistency
> 
> c) eliminate the feature.
> 

My opinion is that if this feature has been broken for two years and no 
one has noticed, then it is time to eliminate the feature.

I can't see a justification for keeping this line number comparison 
shortcut at the same time that Kernel.gsub and related methods are being 
removed.  To put this another way: if -e in Ruby 1.9 allows integer 
range literals in conditionals to be compared against the current line 
number, then I think that -e should also add gsub, sub, split, back in 
to Kernel.

	David