[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Pan-users] Filtering out responses to individuals already filtered
From: |
Duncan |
Subject: |
Re: [Pan-users] Filtering out responses to individuals already filtered out |
Date: |
Wed, 24 Jul 2013 02:09:39 +0000 (UTC) |
User-agent: |
Pan/0.140 (Chocolate Salty Balls; GIT 80384ce /usr/src/portage/src/egit-src/pan2) |
JCA posted on Tue, 23 Jul 2013 18:39:00 -0600 as excerpted:
> I have my filters set up so that posts from a number of individuals are
> deleted automatically before I even get to see them. What I would like
> to do is set up a filter that will do the same to all responses to such
> posts, no matter who is replying. At first I thought this would be
> possible by filtering on the basis of the contents of posts - but Pan
> does not have this capability.
>
> Is it possible to create a Pan filter to do what I am asking, namely, to
> filter out all responses to posts from peoples whose messages are
> already automatically deleted?
First, please tell your mail client not to post in HTML. There's a
reason pan doesn't do HTML, and if you can't respect it in general,
please at least respect it on the pan lists.
For users with a unique string in their message-id header, scoring on
that string in the references header is the likely to be the most
efficient method since that header is in the overview file. Of course
that won't work well if they don't have such a unique message-id header
string and/or for replies from mail clients that don't set the references
header appropriately.
Meanwhile, I haven't actually personally done much with scoring in years,
but IIRC, Heinrich fixed up the scoring so that it would work with any
header; you just have to enter the score manually in the scorefile in
ordered to do it, and of course if the header isn't in the overview file,
the scoring won't apply until the message is actually downloaded, as
that's when pan finally gets all the headers. However, that's still
better than having to manually delete the messages. So if you can find
any header with a unique enough string to score on you can use that, just
understand the compromises you make if it's not an overview header.
However, AFAIK the references header propagating message-id is the only
header that normally appears in all replies.
FWIW, pan's scorefile is slrn's base scorefile format, altho some of
slrn's fancier scorefile features (like include directives, I believe,
and grouping) aren't yet available in pan. Here's the link to their
documentation:
http://www.slrn.org/docs/score.txt
(Note that you can still /sort/ /of/ do grouping -- you simply split the
intended score up between the conditions such that only if they all match
does it get a full score, otherwise the score is modified some by the
parts that do match but not enough to get it into the target category,
presumably ignored, in this case.)
However, given the limitations of the references and message-id headers,
if you can't find a unique string to score on, you basically have to
choose between either auto-act deleting the direct offenders and doing
the score subthread (which works on the references header) entirely
manually when you see replies, OR, not to auto-action delete ignored
posts but instead simply score them ignored, setting pan to show ignored
posts, so then as soon as you see the offending root, you can use score-
subthread (manually, but you can at least use the ignored color-coding to
spot the root) on it and thus kill the entire subthread from there.
[I hope that all makes sense. It seems to here, but I've been up about
30 hours straight due to a weird work schedule and expect to be crashing
as soon as the caffeine wears off, so I'm probably not the best judge of
it making sense or not ATM.]
--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: [Pan-users] Filtering out responses to individuals already filtered out,
Duncan <=