I'm back, change in web archives

I'm back, change in web archives

[Jason Molenda]

I'm back from vacation.

After a week off, I have a huge pile of incoming mail to read, so
naturally I decided to blow it all off for an hour or two and fix a
long-standing TODO for the sourceware site.

The web archives are now a little more resistant to spam harvesters.
E-mail addresses are no longer listed in the subject and author indexes
(just the full name of the poster), and e-mail addresses in message
headers are modified with an expression like s,@, AT ,;s,.com, -dot-
com,; cygwin@sourceware.cygnus.com becomes cygwin AT sourceware.cygnus
-dot- com.

It's ugly, I know.  I'm open to better ideas.  I don't know if it's
worth the trouble of the second ".com, -dot com," substitution.

e-mail addresses in the body of messages are untouched.

The big web archive services on the net like www.egroups.com and
www.mail-archive.com go much farther than I have.  mail-archive.com
omits e-mail addresses in headers altogether.  egroups.com changes the
last character in the local part to a "-" and provides a link that you
can follow to get the correct address if you're a registered member.
(at least that's what it looks like the link does--I'm not a registered
member).


My change is not too effective if the harvester is clever enough to
write a regular expression fashioned to my particular rewrite rules.
I don't want to resort to destructively modifying the addresses like
these other sites have done, unless I really really have to.

Jason // Sourceware Czar
         Free the Software!