From: Mauricio Fernandez <mfp@...>
Date: 2006-02-12T18:09:30+09:00
Subject: Re: Parse error within Regexp

On Sun, Feb 12, 2006 at 11:42:42AM +0900, dblack@wobblini.net wrote:
> >Would something like this make sense?
> >
> >$  ./ruby19 -e '"".scan / ,/'
> >-e:1: syntax error, unexpected ','
> >"".scan / ,/
> >          ^
> >-e:1: unterminated regexp meets end of file
> >                 ========
> 
> I'm not sure the "unterminated string" makes sense as the error
> message, but I don't think it's an unterminated regex either.  It
> seems that Ruby sees the first / as division, so it doesn't consider
> itself to be in mid-regex.  If it thought it was in mid-regex, it
> would presumably keep going and it would find that the regex *is*
> terminated.

I think the second '/' marks the start of the regexp, as seen by
the parser in that case (erroneously of course, but read below):

"".scan / ,/   <----------------------------
          ^                                  \
this is the unexpected ,                  and this the unterminated regexp

Compare that to

$ ruby -e '"".scan / , " '
-e:1: syntax error, unexpected ','
"".scan / , "
           ^
-e:1: unterminated string meets end of file
-e:1: warning: useless use of a literal in void context

The parser is doing its best to keep parsing so it can report further
errors.

Besides, consider this case:

$ ruby19 -e 'puts ARGF.read.scan /[-+]?(\d+(?:\.\d+)?)(?=[^\d])'
-e:1: warning: ambiguous first argument; put parentheses or even spaces
-e:1: unterminated string meets end of file
-e:1: syntax error, unexpected tSTRING_END, expecting tSTRING_CONTENT or tREGEXP_END or tSTRING_DBEG or tSTRING_DVAR

I think the following makes more sense and is less noisy too:

~/src/ruby/ruby.head$ ./ruby19 -e 'puts ARGF.read.scan /[-+]?(\d+(?:\.\d+)?)(?=[^\d])'
-e:1: warning: ambiguous first argument; put parentheses or even spaces
-e:1: unterminated regexp meets end of file
-e:1: warning: parenthesize argument(s) for future version


-- 
Mauricio Fernandez