in reply to Getting Email environment address

I think your best bet is to have them enter their email, and hope for the best. For verification, pattern match for something that looks like a valid email. Since there are many possibilities: foo@bar.(com | org | edu | gov| uk | org.gy), first.last@foo.bar.com, etc, you'll have to be very creative. As a matter of fact, Mastering Regular Expressions has 10 pages dedicated to this problem, and uses it as a "Let's put everything we've learned in the entire book" example. At the very least, you should make sure there's a '@' and at least one '.' in the address.

If you wanna be more picky, you can start a background mailer that sends a test mail to the address. If it bounces, the address is bogus. This'll look a little odd on their end, and I can't guarantee how they'll react, but it should work.

-Logan
"What do I want? I'm an American. I want more."

Replies are listed 'Best First'.
Re: Re: Getting Email environment address
by LAI (Hermit) on Nov 15, 2002 at 14:07 UTC

    I wrote a regexp a little while ago that will validate an email pretty well. It allows for pretty much anything before the '@' except for whitespace and '@', then allows alphanumerics and hyphens in the domain name (with the same rule for n subdomains), then an alphabetical tld. I figured it would be a waste of time to maintain a list of all the tlds, so I just check to make sure the tld has two or 3 letters, or is 'info' or 'museum'. 

    /^[^@\s]+@([a-z0-9\-]+\.)+([a-z]{2,3}||info||museum)$/i

    [download]

    I would very much like to know if anyone can offer improvements on this, as I plan to use it again soon, but I've tried quite a number of garbage addresses and they're stopped. A limitation of this regexp is that characters like ';' and ',' are not stopped in the first part, so someone could theoretically send extra addresses -- but the '@' is stopped. So, someone could theoretically get your program to send mail to local accounts, or just generate errors. I felt this was an okay compromise, but to stop this you could change the regexp to: 

    /^[^@\s;,]+@([a-z0-9\-]+\.)+([a-z]{2,3}||info||museum)$/i

    [download]

    Of course, if you really want to validate the address, you could always do a DNS lookup on the domain name (don't laugh, I know people who've done it), but I can't think of any way to check if the address is valid...


    LAI
    :eof
[reply]
[d/l]
[select]
      foo" "bar@abigail.nl is a valid address, but it will be stopped by your regex. There are a few modules on CPAN that check addresses against RFC 822, for instance RFC::RFC822::Address.

      Abigail

[reply]
[d/l]
[select]

        Thanks Abigail-II; I had no idea that sort of address is valid. Perhaps I should have RTFM!

        Now to read up on RFC 822...


        LAI
        :eof
[reply]

In Section Seekers of Perl Wisdom