[Perl-beginners] confused about a regex

I think you will be less confused if you change the code to

perl -e '$_="abc.eeeee.i";
s/(\.\w+)?$/.out/;
print;'

abc.eeeee.i is actually 'abc' . 'eeeee' . 'i' or 'abceeeeei'. You
might want to look up Perl's behaviour with respect to barewords.

That said the regex replaces either a period followed by one or more
word characters (A-Z, a-z, 0-9, and _) at the end of the string or
nothing at the end of a string with ".out".

$_=abc.eeeee.i;

This is short for:

$_ = 'abc' . 'eeeee' . 'i';

Which is the same as saying:

$_ = 'abceeeeei';


The substitution:

s/(\.\w+)?$/.out/;

Says to match a period (.) followed by one or more 'word' characters (\w+).
This pattern is enclosed in parentheses and this group should match zero or
one time. The pattern is anchored to the end of the string. Because the
pattern will always match, the replacement string '.out' will either be
appended to the string or replace any (\.\w+) pattern at the end of the string.



John

Why is $_=abc.eeeee.i short for
$_ = 'abc' . 'eeeee' . 'i';

Is it b/c each group of characters is a 'token'
including the periods?

abc => token
. => token
eeeee => token
. => token
i => token

From the Perl CD:
the lexical analyzer breaks it down into three tokens:
print, "Hello, world!\n", and the final semicolon/



____________________________________________________________________________________
Sucker-punch spam with award-winning protection.
Try the free Yahoo! Mail Beta.
http://advision.webevents.yahoo.com/mailbeta/features_spam.html

On 3/8/07, oryann9 wrote:
snip
snip

from "Programming Perl 3rd Edition":
bareword
A word sufficient ambigious to be deemed
illegal under use strict 'subs'. In the
absence of that stricture, a bareword is treated
as if quotes were around it.

Basically a bareword is anything that matches /[_a-zA-Z][\w\d]*/ and
does not have enough context around it for the perl interpreter to
decide what it is. For instance, the FILE in the code below matches
the bareword pattern, but the context of open and print let the perl
interpreter know it is actually a file handle*.

open FILE, '>', $file or die "could not open $file: $!";
print FILE "FILE is not a bareword";

But in the case we are talking about the interpreter does not know
what to do with abc, eeeee, or i. They aren't file handles. They
aren't functions** being called with no arguments (because the
interpreter hasn't seen a definition for abc, eeeee, or i yet). So it
just throws quotes around each of abc, eeeee, and i and moves on.

Now all of this begs the question "why do barewords exist at all?".
The short answer is they don't if you use the strict pragma like you
are supposed to. The longer answer is that Perl is multipurpose
language. It supports styles and usages that are outside of the
normal programming scope. One of its many uses is as a commandline
tool. I use it everyday instead of the common SUS*** utilities grep,
cut, sort, etc. I don't sit down and write a file to run through the
interpreter; instead I say things like

cdparanoia -Q 2>&1 | perl -ane 'print $F[2] if $F[0] eq TOTAL'

In that environment I often want to cut down the number of characters
I have to type to get the job done and barewords are one of the things
Perl lets me use to achieve that goal. Barewords also have uses that
the strict pragma doesn't ban. Specifically the ones involving
hashes:

my %hash = (
foo => 'foo is a bareword'
);

$hash{bar} = 'if the key is a bareword I don't have to quote it';
$hash{'not this'} = 'this key contains a non-bareword character, so it
must be quoted';

I hope this has been helpful to you.

* but you should be using bareword like file handles anymore either.
Use IO::File and the "open my $fh, '>', $file" idiom instead.
** This is sort of how use constant works, try this at home
#!/usr/bin/perl

use strict;
use warnings;

sub abc { "abc." }
sub eeeee { "eeeee." }
sub i { "i" }

$_=abc.eeeee.i;

print "$_\n";

*** Single Unix Specification, or what used to be known as POSIX

Yes I understand now. For some reason I missed the
missing quotes in the original post and the word token
came to mind.

$ perl -MO=Deparse foo.plx
BEGIN { $^W = 1; }
use diagnostics;
sub abc {
use warnings;
use strict 'refs';
'abc.';
}
sub eeeee {
use warnings;
use strict 'refs';
'eeeee.';
}
sub i {
use warnings;
use strict 'refs';
'i';
}
use warnings;
use strict 'refs';
$_ = abc() . eeeee() . i();
print "$_\n";
foo.plx syntax OK




____________________________________________________________________________________
Need Mail bonding?
Go to the Yahoo! Mail Q&A for great tips from Yahoo! Answers users.
http://answers.yahoo.com/dir/?link=list&sid=396546091

$_=abc.eeeee.i;

This is short for:

$_ = 'abc' . 'eeeee' . 'i';

Which is the same as saying:

$_ = 'abceeeeei';



Thanks.I never knew that it can write a string like this way.

_____________________________________________________________
FREE Email @ Fadmail.com - http://www.fadmail.com

You probably shouldn't though. It is a carry over from the earlier
days of Perl. Modern Perl should use the strict pragma (which bans
barewords outside of a few special cases).

I have to say - I am totally enamored with regex. Color me
'goober'. I just think that is a beautiful, concise, elegant way to
make a substitution. All of that capability in one short string of
characters... No if, then, else construct. Just - capture what is
there; if it matches a .\w then change it to a .out. If there isn't
a .\w at the end of the string, just append the .out.