Hello Januk,

   On Fri, 7 May 2004 09:58:36 -0700 (07.05.2004 22:58 my local time),
   received Saturday, May 8, 2004 at 10:47:26 +0600,
   you wrote about "Filtering email addresses out of  message body"
   at least in part:

JA> I'm sure you could do it. :-)

JA> The basic format is pretty much the same for all the templates I've
JA> written.
If (and when) I'll use it, I'll prefer using (for|while) cycles from
MyMacros - for simplicity _and_ readability

AL>> [_a-zA-Z\d\-\.\+]+@([_a-zA-Z\d\-]+(\.[_a-zA-Z\d\-]+)+)

JA> That's interesting.  I didn't know that the + character was allowed in
JA> the username portion of an address.
Yes, sure... Read RFC-2822 as single correct source (well, with one
addon - RFC address only SMTP-addressing scheme and doesn't cover X.400,
UUCP transport)

JA> According to that regexp, <[EMAIL PROTECTED]> is a valid e-mail
JA> address.
Yes. If MDA can handle "detail" part of address (after "+" sign), in can
perform special processing of message, otherwise all messages with any
details will be delivered to mailbox, assigned for user lucy (in your
example)

JA>  If so, I wonder why TB doesn't recognize it as such.
Small mistakes always possible

JA> Also, did you know that you could have matched any alphanumeric
JA> character with "\w"?
Yes, but in my form I have regexp
- directly expanded from RFC-2822 specification
- independent from engine-specific regexp implementations (they can know
nothing about POSIX extensions and even character classes)

JA> Just one of those tidbits that might make your life a bit easier at
JA> some point. :-)
I prefer readability and manageability :-) (long life instead of easy
life).

-- 
Best regards,
 Alexander



________________________________________________________

http://www.silverstones.com/thebat/TBUDLInfo.html

Reply via email to