Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
The above filter entry does not match initial articles. I tried multiple versions, including using msgid_last and msgid_only, but could not
achieve what I wanted. I did not spot a pattern what is matched.
Can I somehow see what tin filters to debug this?
Does someone here has a working filter and would share it?
On Mon, 9 Oct 2023 07:55:13 -0000 (UTC), Rene Kita wrote:
Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with
'@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
The above filter entry does not match initial articles. I tried multiple
versions, including using msgid_last and msgid_only, but could not
achieve what I wanted. I did not spot a pattern what is matched.
Can I somehow see what tin filters to debug this?
Does someone here has a working filter and would share it?
This site is now a few years old, but its tin filter may still work.
"The Usenet Improvement Project - Filters Examples 3"
<http://twovoyagers.com/improve-usenet.org/filters_ex3.html>
issdr wrote:
Rene Kita wrote:
msgid=.*@googlegroups.com>
try this:
msgid=.*@googlegroups\.com
nevermind. you should check tin's wildcard variable;
however, this should work in any case:
msgid_only:@googlegroups.com
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
msgid=.*@googlegroups.com>
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
Rene Kita wrote:
msgid=.*@googlegroups.com>
try this:
msgid=.*@googlegroups\.com
To quote the important part from that site:
group=*
case=0
score=kill
msgid_only=googlegroups
I tried that and articles with a Message-Id containing googlegroups are
not killed.
Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
The above filter entry does not match initial articles. I tried multiple versions, including using msgid_last and msgid_only, but could not
achieve what I wanted. I did not spot a pattern what is matched.
Can I somehow see what tin filters to debug this?
Does someone here has a working filter and would share it?
Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
The above filter entry does not match initial articles. I tried multiple versions, including using msgid_last and msgid_only, but could not
achieve what I wanted. I did not spot a pattern what is matched.
Can I somehow see what tin filters to debug this?
Does someone here has a working filter and would share it?
I do not want to filter followups to posts from GG users.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
Can I somehow see what tin filters to debug this?
As always, it was a PEBKAC. I had another rule that also matched.
Apparently, if you have a msgid and a from in one rule those are not
ANDed but ORed.
Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with '@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
Does someone here has a working filter and would share it?
Can I somehow see what tin filters to debug this?
If tin is compiled with debugging support and started with "-D 2"
it will create $TMPDIR/FILTER (details about the filter rules) and $TMPDIR/ARTS (status of the articles after applying the rules), but
those files are hard to read without having some knowledge about the data-structures used. there is no simply "article killed due to the
following rule" output available.
As always, it was a PEBKAC. I had another rule that also matched.
Apparently, if you have a msgid and a from in one rule those are not
ANDed but ORed.
yes, there is currently no AND (but you could use scores with sum up to score_limit_kill/score_limit_select only if all the rules are matched as
some kind of a workaroud.
Rene Kita <mail@rkta.de> wrote:
Due to recent events I finally tried to filter Google Groups users with
tin. I do not want to filter followups to posts from GG users.
AFAIU GG posters have a Message-ID which always ends with
'@googlegroups.com'.
I have the following regex in my filter file:
group=*
case=0
score=-25
msgid=.*@googlegroups.com>
A few years ago tin made changes to their filter logic and I
found this works for me:
group=*
case=0
score=kill
msgid_last=*<*@googlegroups.com>*
Urs recommended to use msgid_only. What difference does msgid_last make
here? I guess some replies will be killed, too.
In <ug2utl$10f6d$2@dont-email.me> on Tue, 10 Oct 2023 09:36:23,
Rene Kita wrote:
Urs recommended to use msgid_only. What difference does msgid_last make
here? I guess some replies will be killed, too.
[tin(5)]
| ${TIN_HOMEDIR:-"$HOME"}/.tin/filter
[...]
| msgid_last
| Match against ’’Message-ID:’’ and last ’’References:’’
| entry only. The matching type used (wildmat or regex) deâ€
| pends on the setting of the tinrc variable wildcard.
|
| msgid_only
| Match against ’’Message-ID:’’. The matching type used
| (wildmat or regex) depends on the setting of the tinrc
| variable wildcard.
so with msgid_last direct responses would be matched too.
JFTR I've added the following to tin(5) to clarify things:
#v+
== modified file 'doc/tin.5'
--- old/doc/tin.5 2023-07-31 06:03:39 +0000
+++ new/doc/tin.5 2023-10-09 19:57:53 +0000
@@ -759 +759,3 @@
-lines beginning with a '#' are ignored.
+lines beginning with a '#' are ignored. Multiple matching-rules in the same +scope are all applied individually and not only when they all hit at the +same time (logical OR not AND).
@@ -879,0 +882,9 @@
+
+comment=downscore postings in nsr from google base on
+comment=message-id or path, if both hit they end up with
+comment=a score of -66 which is enough to mark them killed +group=news.software.readers
+case=1
+score=-33
+path=\\.googlegroups\\.com!not-for-mail$
+msgid_only=@googlegroups\\.com>
#v-
Maybe you could clarify this too. For me it was not obvious that 'last References entry' means direct responses.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 482 |
Nodes: | 16 (0 / 16) |
Uptime: | 70:43:58 |
Calls: | 9,571 |
Calls today: | 2 |
Files: | 13,663 |
Messages: | 6,142,257 |