1519_SPAM

You never had any friends, and don't really want any anyways, and so you have decided to collect email addresses from web pages for direct e-mail advertising.

The text delivered to a web browser is usually marked up HTML, which may contain email addresses of the form:

user@server

¨ Bothuserandserverare of the formalpha.numeric.with.dots. Byalpha.numeric.with.dots, we mean a sequence of one or more characters which are alphabetic ( A-Z, a-z ), numeric ( 0-9 ), hyphens (-), underbars (_) and/or periods (.), with the following restrictions on periods:

n The sequence neither starts nor ends with a period.

n No periods are adjacent.

¨ Email addresses are preceded by the beginning of the file, or some character other than a letter ( A-Z, a-z ), digit ( 0-9), hyphen (-), or underbar (_).

¨ Email addresses are succeeded by the end of the file, or some character other than a letter ( A-Z, a-z ), digit ( 0-9 ), hyphen (-), or underbar (_).

¨ If the scanned text contains a sequence of the form

first@second@third

Then the output should containfirst@secondandsecond@thirdas email addresses. In a longer run, each pair split by an @-sign should appear as an email address in the output.

The point of this problem is to extract and record the email addresses embedded in other text.

The input file will contain zero or more lines of ASCII text.

Other than the standard leader and trailer, the output file has each email address found in the input file in the order it was found (duplicates not removed).

bob@banks.com wrote:
What does x=7 mean for this problem?  For
example,

..a@a@aa@aaa@aaa..a@a@aa@aaa@aaa..a@a..@a...a@..@..

this scrolling @-example from jim@jones.com

bob@banks.com
a@a
a@aa
aa@aaa
aaa@aaa
a@a
a@aa
aa@aaa
aaa@aaa
a@a
jim@jones.com