[Users] Filtering spam difference - auto v. manual

[Pierre Fortin]

Hi Jeremy,

Although it doesn't address my query, your point is a valid one, so I'm
putting it back on the ML.

I originally had regexp unchecked; but since I include ">" in the
pattern, that ensures:
1. it's [part of] TLDs I want to block 
2. it catches anything ending in those patterns; all of which are most
likely spam.
It would be nice if matcherrc included a comments section; but...  :)

So, in your example of abid vs bid, I can't conceive of valid msgs from
*[pattern]> TLD {bid,abid,bestbid,highbid,...} that I wouldn't consider
spam. 

There are so many TLDs now, it's no longer reasonable to use global
block|allow lists.  All I can do is block the new ones that hit my inbox,
and let spamd build its DB after which I can remove some old patterns;
but we know how often cleanups occur... LOL

Cheers,
Pierre

On Sun, 29 Oct 2017 10:27:17 +0000 Jeremy Nicoll wrote:

>On Sun, 29 Oct 2017, at 02:03, Pierre Fortin wrote:
>
>[snip]
>
>I don't know the answer to your question, but I think you 
>have a small problem with some of your rules.  Something
>like
>
>>     from regexpcase ".bid>" |   
>
>is looking for a 'From' value ending not in  dot bid >    as
>you possibly think, but instead
>
>   any single character, then bid, then >
>
>It will be 'working' because the 'any single character' 
>part of that test does match on the dot present in such 
>mails.
>
>But if a mail came from eg   foo.abid>     that test would
>match as well with the dot in the regex pattern matching
>the "a" before "bid".
>
>Your pattern, if you want it to match on a literal dot, 
>should have a backslash before the dot, ie 
>
>     from regexpcase "\.bid>" 
>