Re: a match question

[John W. Krahn]

Shawn H Corey wrote:
> Jim Gibson wrote:
> > On 11/24/09 Tue  Nov 24, 2009  4:42 PM, "Orchid Fairy   (À¼»¨ÏÉ×Ó)"
> > <practicalperl@xxxxxxxxx> scribbled:
> >
> > > On Wed, Nov 25, 2009 at 8:38 AM, Shawn H Corey <shawnhcorey@xxxxxxxxx> wrote:
> > >
> > > > It seems to be picking up an extra empty string after a non-zero length
> > > > match.
> > > Thanks John and Shawn.
> > > Yup what let me be confused is that, why there is an additional empty
> > > string there?
> > You get 4 matches because there are four places in the string 'aXXXb' that
> > can be matched by 'X*'. Note that a "place" in this case is a position
> > between characters. The places are:
> >
> > 1. before the a
> > 2. after the a
> > 3. before the b
> > 4. after the b
>
> Funny, I get this on my machine:
>
> $ perl -le '$_="aXXXb"; print "one word is $1 at ", pos while(/(X*)/g);'
> one word is  at 0
> one word is XXX at 4
> one word is  at 4
> one word is  at 5
>
> That is:
> 1. before the a
> 2. between the a and b
> 3. before the b
> 4. after the b
>
> The problem is that the third one should not be there.

Yes it should.  Perhaps you are confused because pos() reports the 
current position *after* the match has occured?  A zero width match has 
the same position at the beginning and end however the 'XXX' string has 
a different position at the beginning and end.  Try using $-[0] for the 
beginning of the match and $+[0] for the end of the match.

$ perl -le '$_="aXXXb"; print "one word is $1 at $-[0] to $+[0]" while 
/(X*)/g;'
one word is  at 0 to 0
one word is XXX at 1 to 4
one word is  at 4 to 4
one word is  at 5 to 5



John
--
The programmer is fighting against the two most
destructive forces in the universe: entropy and
human stupidity.               -- Damian Conway