jeanluca has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks

Here is a script which replaces the last part of an URL 
#! /usr/bin/perl -l

my $url = "http://example.com/user/1234/" ;
$url =~ s![^/]+(?=/$)!_bla_! ;
print "URL = $url" ;

[download]
this works. However when I change the regex to 
s![^/]+(?=/)$!_bla_!

[download]
it doesn't!
Any suggestions why ?

cheers
LuCa

UPDATE: thnx a lot, I understand it now!!

Replies are listed 'Best First'.
Re: regex issue (positive look-ahead)
by kennethk (Abbot) on Mar 10, 2010 at 15:28 UTC
    The difference between the two regexes is the placement of your $ metacharacter, so that should be your red flag. By moving it you are changing your trailing anchor. The first version says:
    1. Match at least one character that is not a slash
    2. It should be followed by (non-consuming):
      1. a slash
      2. the end of string

    The new version says:

    1. Match at least one character that is not a slash
    2. It should be followed by (non-consuming):
      1. a slash
    3. The last consumed element should be followed by the end of string

    In your second regex, the positive look-ahead assertion and the end of string anchor directly conflict. Obligatory documentation refs: perlre, perlretut.

[reply]
[d/l]
Re: regex issue (positive look-ahead)
by ikegami (Patriarch) on Mar 10, 2010 at 15:36 UTC
    zero-width positive lookahead. s![^/]+(?=/)$!_bla_! finds non-slashes, looks ahead to make sure they are followed by a slash, then checks if they are followed by a newline or end of string. That can't happen. You can't have something that's followed both by a slash and by a newline, and you can't have something that's followed both by a slash and nothing.
[reply]
[d/l]
Re: regex issue (positive look-ahead)
by rubasov (Friar) on Mar 10, 2010 at 15:42 UTC
    In short: because the lookahead you're using is a zero-width assertion.

    In a more explanatory style: your second regex means something like this: match a url part not containing slashes, followed by a slash (but do not include this slash into to the currently matched substring), and followed by the end of the string. 

    url part not containing slashes -|- followed by a slash
                                 |- followed by the end of the string

    [download]
    Those followed by clauses should be fulfilled at the same time, however your url does not contain such a substring.

    update: corrected a negation error above, thanks ikegami.

[reply]
[d/l]
[select]
Re: regex issue (positive look-ahead)
by Anonymous Monk on Mar 10, 2010 at 15:43 UTC 
    use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new( qr![^/]+(?=/$)! )->explain;
__END__
The regular expression:

(?-imsx:[^/]+(?=/$))

matches as follows:

NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  [^/]+                    any character except: '/' (1 or more times
                           (matching the most amount possible))
----------------------------------------------------------------------
  (?=                      look ahead to see if there is:
----------------------------------------------------------------------
    /                        '/'
----------------------------------------------------------------------
    $                        before an optional \n, and the end of
                             the string
----------------------------------------------------------------------
  )                        end of look-ahead
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------

    [download]
    use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new( qr![^/]+(?=/)$! )->explain;
__END__
The regular expression:

(?-imsx:[^/]+(?=/)$)

matches as follows:

NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  [^/]+                    any character except: '/' (1 or more times
                           (matching the most amount possible))
----------------------------------------------------------------------
  (?=                      look ahead to see if there is:
----------------------------------------------------------------------
    /                        '/'
----------------------------------------------------------------------
  )                        end of look-ahead
----------------------------------------------------------------------
  $                        before an optional \n, and the end of the
                           string
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------

    [download]
[reply]
[d/l]
[select]
Re: regex issue (positive look-ahead)
by Anonymous Monk on Mar 10, 2010 at 15:10 UTC
[reply]
[d/l]

Back to Seekers of Perl Wisdom