Gossamer Forum: General: Perl Programming: Newbie

Jul 29, 2002, 2:51 PM

shaan

Novice (13 posts)

Jul 29, 2002, 2:51 PM

Post #1 of 26

Shortcut

Newbie - brain on fire!

Hi guys,
I am new to programming and perl. Am learning how to create feedback forms. I seem to get stuck at email validation. What am I doing wrong? Can someone please help me. Unsure

Thanks a lot. Here is how my script reads

#Validate input of correct email address

$good_email = 0;

@check_email = split (//, $email);
foreach $letter (@check_email)

{
if ($letter eq "@")
{
$good_email = 1;
last;
}
}

print("Content-Type: text/html\n\n");
print "<HTML>\n<HEAD>\n";

if (!$good_email)
{
print "<TITLE>Invalid email address</TITLE>\n\<BODY>\n";
print "<H2>Invalid Email Address detected</H2>";
print "<p>Please provide correct email address. Go back to the form</p>\n";
}
else
{
print "<TITLE>Thank You</TITLE>\n<BODY>\n";
print "<P>Thank you for filling out the form. We will contact you soon at $email</P>";
}

Jul 29, 2002, 4:59 PM

Paul

Veteran (19537 posts)

Jul 29, 2002, 4:59 PM

Post #2 of 26

Shortcut

Re: [shaan] Newbie - brain on fire! In reply to

If you look down about 10 threads you'll see:

reg ex to validate email address

Jul 29, 2002, 6:30 PM

widgetz

Veteran (2260 posts)

Jul 29, 2002, 6:30 PM

Post #3 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

or.. you can stay right here.. and ignore paul.

remove this..

$good_email = 0;

@check_email = split (//, $email);
foreach $letter (@check_email)

{
if ($letter eq "@")
{
$good_email = 1;
last;
}
}

replace

if (!$good_email) {

with

if ($email !~ /^[\w_-\.+]+@[\w_-\.]\.\w+$/) {

not sure if those are all the characters that can go in an email.. i think it is.. oh well.

Jul 30, 2002, 7:46 AM

shaan

Novice (13 posts)

Jul 30, 2002, 7:46 AM

Post #4 of 26

Shortcut

Re: [widgetz] Newbie - brain on fire! In reply to

Revised script...

@check_email = split (//, $email);
foreach $letter (@check_email)
{
if ($letter eq "@")
{
$good_email =1;
last;
}
}

print("Content-Type: text/html\n\n");
print "<HTML>\n<HEAD>\n";

if ($email !~ /^[\w_-\.+]+@[\w_-\.]\.\w+$/) {
{
print "<TITLE>Invalid email address</TITLE>\n\<BODY>\n";
print "<H2>Invalid Email Address detected";
print "<p>Please provide correct email address</p>\n";
}
else
{
print "<TITLE>Thank You</TITLE>\n<BODY>\n";

QUESTION
1. Don't I have to assign an initial value to $good_email?

2. Can I use "==" instead of "eq" in if ($letter..)?

Jul 30, 2002, 7:53 AM

Paul

Veteran (19537 posts)

Jul 30, 2002, 7:53 AM

Post #5 of 26

Shortcut

Re: [shaan] Newbie - brain on fire! In reply to

You should be aware that Jerry's code will match:

______________@______________.__________

.....................@..........................._

-----------@____________._

...and quite a few other combo's. If you want something reliable you should use the Email::Valid perl module.

http://search.cpan.org/search?dist=Email-Valid

Jerry's code is also longer than it needs to be. He has used a character class ( [ ] ) including \w and _ but \w already includes underscores so the _ is redundant.

>>
QUESTION
1. Don't I have to assign an initial value to $good_email?

2. Can I use "==" instead of "eq" in if ($letter..)?
<<

You don't need to know the answer to those questions as you shouldn't be using that code :)

Last edited by:

Paul: Jul 30, 2002, 7:54 AM

Jul 30, 2002, 8:25 AM

shaan

Novice (13 posts)

Jul 30, 2002, 8:25 AM

Post #6 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

Thanks so much guys (P$W)!
It works!!

Thanks for your help.

Jul 30, 2002, 9:28 AM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 9:28 AM

Post #7 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

oh wow.

you're right. lol.

it is possible to have periods in yoru emails.. heres another.

/^[^a-z0-9]+([_-\.]{1}[^a-z0-9]+)*?\@[^a-z0-9]+([_-\.]{1}[^a-z0-9]+)*?\.[a-z0-9]{2,3}$/i

oof. correct that if you see anything wrong with it. it's quite long, but i think it does the job for most situations.

--
jsu

Jul 30, 2002, 9:35 AM

Paul

Veteran (19537 posts)

Jul 30, 2002, 9:35 AM

Post #8 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

/^[^a-z0-9]+([_-\.]{1}[^a-z0-9]+)*?\@[^a-z0-9]+([_-\.]{1}[^a-z0-9]+)*?\.[a-z0-9]{2,3}$/i

Hmm you don't need to escape @'s but more importantly it is poorly formatted (you need to escape - inside [ ]) and won't match basic emails like pwilson@aol.com as you are blocking a-z and 0-9

Jul 30, 2002, 9:54 AM

Alex

Administrator (9387 posts)

Jul 30, 2002, 9:54 AM

Post #9 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

I would give up trying to come up with a regex to match an email, it's a waste of effort. Even the four page regex posted earlier is not going to match some emails that are deliverable (it only matches rfc compliant emails, however many emails that are not rfc compliant can still get delivered). If your goal is to test that the email is valid, then you need to send a validation email to the user and have them click on a validation link. If your goal is to catch typos, then \S@\S+\.\S+ is fine.

Cheers,

Alex
--
Gossamer Threads Inc.

Jul 30, 2002, 1:25 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 1:25 PM

Post #10 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

that's what i get for doing things in the early morning..

isn't it clear that you don't block the a-z0-9.. thought you'd catch that. also.. i tend to backslash anything that is used for variables in perl. @ $ % { }.. it's just a thing i do. does the compiler care?

/^[a-z0-9]+([_-\.]{1}[a-z0-9]+)*\@[a-z0-9]+([_-\.]{1}[a-z0-9]+)*\.[a-z0-9]{2,3}$/i

--
jsu

Last edited by:

Seto Kaiba: Jul 30, 2002, 1:29 PM

Jul 30, 2002, 1:30 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 1:30 PM

Post #11 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

>>
does the compiler care?
<<

Yes, well it does regarding -

Last edited by:

Paul: Jul 30, 2002, 1:31 PM

Jul 30, 2002, 1:58 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 1:58 PM

Post #12 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

/^[a-z0-9]+([_\-\.][a-z0-9]+)*\@[a-z0-9]+([_\-\.][a-z0-9]+)*\.[a-z0-9]{2,3}$/i

So we agree this works now?

--
jsu

Last edited by:

Seto Kaiba: Jul 30, 2002, 2:01 PM

Jul 30, 2002, 4:37 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 4:37 PM

Post #13 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

Well I agree it compiles, but it isn't very good :)

Try running this...

Code:
#!/perl/bin/perl 
#========================================================== 

   use strict; 

   main(); 

#========================================================== 

sub main { 
#---------------------------------------------------------- 
# Oops! 

    print "Content-type: text/plain\n\n"; 
    foreach (qw/pwilson@wiredon.net  
                pwilson@aol.com  
                pwilson@aol.co.uk  
                test@123-domain.com  
                test@123-domain.co.uk  
                foo@bar.info 
                me@here.info 
                foo@123-domain.ws 
                i_m@valid.com 
                _imvalid@here.com/) { 
        print /^[a-z0-9]+([_\-\.][a-z0-9]+)*\@[a-z0-9]+([_\-\.][a-z0-9]+)*\.[a-z0-9]{2,3}$/i ? "$_ is OK\n" : "$_ is NOT OK\n"; 
    } 
}

Last edited by:

Paul: Jul 30, 2002, 4:38 PM

Jul 30, 2002, 4:44 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 4:44 PM

Post #14 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

_imvalid@here.com is valid??

err.. if so.. then just add _ into the first bracket.. but i'm not even allowed to create _email@jsu07.com on my server

--
jsu

Jul 30, 2002, 4:47 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 4:47 PM

Post #15 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

>>
but i'm not even allowed to create _email@jsu07.com on my server
<<

I can, I didn't bother to check if it was valid or not but anway the regex still won't match info domains.

>>
err.. if so.. then just add _ into the first bracket..
<<

I wouldn't ;)

Like Alex said, i'd give up.

Last edited by:

Paul: Jul 30, 2002, 4:47 PM

Jul 30, 2002, 4:54 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 4:54 PM

Post #16 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

didn't think of 4 letter tlds.

/^_?[a-z0-9]+([_\-\.][a-z0-9]+)*\@[a-z0-9]+([_\-\.][a-z0-9]+)*\.[a-z0-9]{2,4}$/i

if you want to be critical.

"give up"..

never. that's something i've never done. and i think using 1 line surely beats a whole module or library.

--
jsu

Jul 30, 2002, 5:00 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 5:00 PM

Post #17 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

Like Alex mentioned without writing something huge you won't even get close to writing something decent.

I'd prefer to use \S+@\S+\.\S+ rather than using yours which has the potential to reject many valid addresses.

Jul 30, 2002, 5:11 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 5:11 PM

Post #18 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

eh. I've just looked at some asp source codes in the msdn.

/\w+((-\w+)|(\.\w+)|(\_\w+))*\@[A-Za-z0-9]+((\.|-)[A-Za-z0-9]+)*\.[A-Za-z]{2,5}/

is what they use. however, i can see where that may go wrong.

the problem with using \S+@\S+\.\S+ is you can get.. @@@.@ and many other combonations of non-whitespace emails.

at least make a check for a single @ and filter out the other invalid characters.

--
jsu

Jul 30, 2002, 5:12 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 5:12 PM

Post #19 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

>>
the problem with using \S+@\S+\.\S+ is you can get.. @@@.@ and many other combonations of non-whitespace emails.

at least make a check for a single @ and filter out the other invalid characters.
<<

I'd prefer it to accept silly emails than reject valid ones.

Jul 30, 2002, 5:23 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 5:23 PM

Post #20 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

my point is.. rather than using \S+ use something else..

[\w\.]+@[\w\.]+\.[A-Za-z]+

will even work better. just figure out what all those valid characters are.. i think \w\. are the only ones. \w matches _ and -, so A-Z a-z 0-9 _ - and . are all there.

--
jsu

Jul 30, 2002, 5:26 PM

Paul

Veteran (19537 posts)

Jul 30, 2002, 5:26 PM

Post #21 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

You aren't grasping this are you :) ...as soon as you start trying to get accurate you'll end up blocking valid emails, you either need to be very general or very precise with your regex.

Do you really think Email::Valid (on CPAN) would be about 50 lines long if it was as simple as you are trying to make out.

Jul 30, 2002, 6:39 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 6:39 PM

Post #22 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

Code:
/^[^\s<>@"\:]+@([\w\-]+\.)*([\w\-]+)\.[a-z]{2,5}$/i && $2 !~ /(__|--)/ 

prove me wrong. please. i tested it on every email address possible and impossible.. i turned on my wildcards and made myself a new email address.

~`!#$%^&*-_+={}[]|;,.,'?/@jsu07.com

it works nicely. ------------.jsu07.com can be used too.. it's not a one-liner.. but it WILL validate email addresses.

think about it.. email address validation can't be too hard. if its possible to validate a domain name.. then you've already got 2/3s of the entire thing. the username can consist of almost anything.. and with that code, i look for the invalids.

i derived it from putting together a longer subroutine originally.. and went through it to shorten the entire thing. the original subroutine i made checked everything 1by1 split up the email address and then i just condensed it (hence all the edits)

--
jsu

Last edited by:

Seto Kaiba: Jul 30, 2002, 7:38 PM

Jul 30, 2002, 8:26 PM

Seto Kaiba

User (105 posts)

Jul 30, 2002, 8:26 PM

Post #23 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

my bad. missed ; and ,

anyways..

/^[^\s<>@"\:;,]+@([\w\-]+\.)*([a-z0-9\-]+)\.[a-z]{2,5}$/i

changed a bit.. because internic says you can't register names with an underscore and also.. no punctuations except for a -..

also...

#!perl
use strict;
use Email::Valid;

# ~`!#$%^&*-_+={}[]|.'?/@jsu07.com

my $email = qq~\~`!#\$%^&*-_+={}[]|.'?/\@jsu07.com~;
print $email . " " . (Email::Valid->address ($email) ? "is valid.\n" : "is not valid.\n");

and finally.. the output..

~`!#$%^&*-_+={}[]|.'?/@jsu07.com is not valid.

of course.. in my inbox..

Envelope-to: "~`!#$%^&*-_+={}[]|.'?/"@jsu07.com
X-Sender: email@mail.jsu07.com
X-Mailer: QUALCOMM Windows Eudora Version 5.1.1
Date: Tue, 30 Jul 2002 20:22:22 -0700
To: ~`!#$%^&*-_+={}[]|.'?/@jsu07.com
From: Jerry Su <email@jsu07.com>
Subject: Test

test

point is.. i'm just as right as Email::Valid.

--
jsu

Last edited by:

Seto Kaiba: Jul 30, 2002, 8:35 PM

Jul 31, 2002, 2:51 AM

Paul

Veteran (19537 posts)

Jul 31, 2002, 2:51 AM

Post #24 of 26

Shortcut

Re: [Seto Kaiba] Newbie - brain on fire! In reply to

>>
~`!#$%^&*-_+={}[]|.'?/@jsu07.com is not valid.

of course.. in my inbox..

Envelope-to: "~`!#$%^&*-_+={}[]|.'?/"@jsu07.com
<<

Email::Valid is right...those characters aren't valid..they may send, but aren't valid.

Last edited by:

Paul: Jul 31, 2002, 3:47 AM

Jul 31, 2002, 7:52 AM

Seto Kaiba

User (105 posts)

Jul 31, 2002, 7:52 AM

Post #25 of 26

Shortcut

Re: [Paul] Newbie - brain on fire! In reply to

you got to ignore those quotes.. eudora adds those in.

also.. email::valid will say valid to domains with exclamation marks and underscores in them. when have you ever seen one of those.

the email address is valid.. try sending an email to it. won't bounce.

--
jsu

Last edited by:

Seto Kaiba: Jul 31, 2002, 8:05 AM

Newbie - brain on fire!

(Page 1 of 2)

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by:

Last edited by: