From: Michael Campbell <michael_s_campbell@...>
Date: 2003-08-30T04:15:42+09:00
Subject: Re: Bug? (1.8.0)

--- Hal Fulton <hal9000@hypermetrics.com> wrote:
> Erik Veenstra wrote:
> > Ruby 1.8.0 on win32:
> > 
> > 1:  ruby -e 'p "#abc#def#".gsub(/#$/, " ")'
> >     -e:1: unterminated string meets end of file
> >     -e:1: syntax error
> > 
> > 2:  ruby -e 'p "#abc#def#".gsub(/#/, " ")'
> >     " abc def "
> > 
> > 3:  ruby -e 'p "#abc#def#".gsub(/\#$/, " ")'
> >     "#abc#def "
> > 
> > What's wrong with the first one? I expected the first one to
> > behave like the third one.
> 
> That's interesting and syntactically subtle.
> 
> Here are the facts the explain this:
> 1. Regexes can have interpolated values just like strings
> 2. When an expression is a single name and that name starts
>     with a special symbol like @, @@, or $, the braces can
>     be omitted
> 3. $/ is one of the ugly special variables
> 4. So #$/ is interpreted as inserting the value of $/ into
>     the regex
> 5. The regex then consumes the rest of the line
> 6. The interpreter apparently does not distinguish between
>     strings and regular expressions when reporting that
>     something is unterminated.
> 
> If there is a bug at all, it is #6 (misleading error message).


Wow, I'd've never gotten that one, and I'm usually pretty good at
catching stuff like that.

I don't think I'd call it a bug, but to my eyes, #2 seems like a
design "whodathunkit".  Mind you, I'm really quite against poetry
mode and a lot of other seemingly "magic" things the interpreter does
(if I want interpolation in a string, I'll tell it so, in this case).

__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com