🚀 go-pugleaf

Like everyone else, I cannot stand downloading some binary which someone has
posted in a discussion group (grrr...).  However, sometimes a discussion is
spawned from the initial complaints added to the thread.

You see, to avoid the binaries I have the lines

Lines:*[0-9][0-9][0-9][0-9]*
Lines:*9[0-9][0-9]*

Which work pretty well, but also:

Subject:*[0-9]/[0-9]*
!Subject:*[0-9]/[0-9][0-9]/[0-9]*
!Subject:*[0-9]/[0-9]/[0-9]*

because many binaries have a (1/1) in the subject.  The otehr two lines are
to stop subject lines with dates being killed.

Howvere, I didn't want to miss any discussion, so I added:

!Subject:*Re:*

This worked great, until today, when someone posted the (0/16) !, then followed
up with the 16 parts, so each part had Re: at the beginning :(

Now, what I am proposing is a change to the processing of the KILL file.

Each line is processed in turn -- if it matches then a flag is set, then it
carries on to the next line, which will set/unset the flag if it matches
(depending on whether it has a '!' at the start).

Thus if I put the Lines: lines at the end, they would always work -- overriding
the Re: line.

Is this feasible?  I have used this system in programs of my own and it has
worked well, allowing complex "decisions" to be made.
-- 
* TQ 1.0 * The 'Just So Quotes'.
"I'm a Leo.  Leos don't believe in this astrology stuff."
    	-- Tom Neff