Thanks Philip for sharing this excellent piece of information.
Cheers,
Parag
On Tue, Dec 22, 2009 at 8:17 PM, Philip Potter <philip.g.pot...@gmail.com>wrote:
> 2009/12/22 Parag Kalra <paragka...@gmail.com>:
> > You can try following:
> >
> > $_ =~ s/^\.(\s)+//g;
>
> This isn't quite right.
>
> There are two ways in which you might use this substitution: either $_
> will contain a single line, or it will contain multiple lines. The
> single line case might look something like this:
>
> while (<>) {
> s/^\.(\s)+//g;
> print;
> }
>
> In this case, however, the /g modifier is redundant and confusing,
> since you only want the regex to match once at the start of each line.
> Take the /g modifier off.
>
> The multiple line case might look like this:
>
> $_ = do {local $/; <>;}; # but better to use File::Slurp
> s/^\.(\s)+//g;
> print;
>
> This code is broken. The problem is that ^ matches start-of-string,
> not start-of-line. Therefore, even with the /g modifier the regex can
> only match at the beginning of the string, and so will only match
> once.
>
> To change ^ to match start-of-line, you need the /m modifier:
>
> s/^\.(\s)+//gm;
> print;
>
> Perl Best Practices makes the interesting recommendation that all
> regexes should use the /m modifier, because having ^$ match line
> boundaries seems to be what most programmers expect and seems to be
> useful much more often. If you still need to match start and end of
> string you can use \A and \z.
>
> Phil
>
