XPost: alt.comp.os.windows-11
"Alan K." <
[email protected]> wrote:
On 8/2/25 11:10 PM, VanguardLH wrote:
PBS Reporting <[email protected]d> wrote:
Path: uni-berlin.de!fu-berlin.de!2.eu.feeder.erje.net!feeder.erje.net!newsfeed.bofh.team!paganini.bofh.team!not-for-mail
^^^^^^^^^^^^^^^^^^
From: PBS Reporting <[email protected]d>
\_________ NYMSHIFTER ____________/
Message-ID: <106m003$1b622$[email protected]>
Could test on this -----\__________________/
<snipped off-topic political spam>
I gave up on paganini.bofh.team. I hide posts that originate from
there. Below is the filter I use to hide those posts.
# Has registered (login) and unregistered (no login) access but no
# headers to identify access type. Paganini becomes the new Google
# Groups for trolls.
!setcolor(olive;yellow),ignore,markread
Header {(?-s)^Path: \S+!paganini\.bofh\.team(!tor-network)?(!not-for-mail)?$}
The Header line is the filter testing on the injection node of the PATH
header. While clients can use their own Message-ID value, clients don't
get to touch the PATH header, so abusers can't lie there.
To test on the PATH header requires downloading more than just the
overview headers. PATH is not an overview header. The filter looks for
the injection node in PATH to show the source in paganini. I don't
delete unwanted article. I hide them, so I could see them if I choose.
Also, deleting them can break the chain of tracking subthreads.
Deleting an article could result in subthreads appearing as new threads.
I configure my client to apply the Ignore flag to child articles. I
don't want to see replies to article I'm hiding.
Rather than the PATH header, I could test on the Message-ID (MID) header
since its right-token identifies paganini. However, if a client
specified a value for MID, the server is supposed to not override with
its own value. Since clients can specify the value, it is not a
reliable header on which to test. While PATH is a non-overview header
(which requires you download the full article, or use a client that
support XPAT *and* the server also supports it), MID is an overview
header, so you should be able to filter without having to download full
articles.
I don't need to colorize those posts, but I hide posts instead of delete
them. My default view is Hide Ignored, but occasionally I switch to
View All, and would like colors to aid in why posts were hidden.
Yeah, you end up tossing all good posters using paganini along with the
bad posters. paganini has unregistered (unauthenticated) access, and
that spurs the trolls to abuse paganini just like they abused Google
Groups and AIOE. paganini, I believe, also has registered
(authenticated) access, but you'll have to try to find its operator to
figure out the setup (he sometimes shows up in alt.free.newsservers).
His placeholder web site is worthless. No information there. I never
figured out if there was any difference in headers to identify posters
using his unregistered access versus his registered access, so I just
hid all of paganini, because of all the aholes abusing it.
Google Groups got so bad with spammers, trolls, malcontents, forgers,
peuriles, and uber-boobs that I filtered them out; see:
http://twovoyagers.com/improve-usenet.org/
AIOE became just as bad (died in Jan 2023). Now paganini is the new
Google Groups. *UN*registered free Usenet providers attract the worst
posters. There is no account to punish abusers. Eternal-September is a
registered free Usenet provider: there is an account to punish by
suspending or killing it. Not a huge hurdle to create another account
to then abuse, but it does present a hurdle.
I like your idea, but can it be done with Thunderbird? And can you
say it in 5 words or less, my eyes started blurring this early in the morning. If you use TB, the easy thing is to extract the filter from msgFilterRules.dat and paste it here. THKS
"He don't know me very well, do he?" (Bugs Bunny). Trying to wrangle
Tbird to test on non-overview headers isn't easy nor quick.
Clients normally only test on overview headers, so that's all they
retrieve for new messages. Not until you select a message does it get
fully downloaded which includes all headers and body. You need to
configure Tbird to download full messages. As I recall, you go into the Synchronization & Storage settings.
I suspect the PATH header is not one of those listed in Tbird's rules.
You have to define the custom header, and then you can test on it. I
found:
https://kb.mozillazine.org/Custom_headers
Even after defining a custom header for PATH, I'm not sure you can
define an adequate filter in Thunderbird since it does not support
regular expressions (regex). You want to test only on the injection
node in PATH. Nodes get prepended to PATH as it passes through each
server, so you see:
PATH: nodeN!...!node3!node2!node1
last first
My paganini regex tests on:
(?-s)^Path: \S+!paganini\.bofh\.team(!tor-network)?(!not-for-mail)?$
The (?-s) makes sure the filter only tests on 1 physical line to prevent
side effects should it extend into other lines for other headers. The
PATH header cannot be wrapped to extend to more lines. Other headers
may wrap by having continuation lines (start with whitespace versus
header lines where the header name must start in column 1).
^Path has the ^ to anchor the substring to the left end of the string;
i.e., PATH starts in column 1. Header names are delineated from their
string value by whitespace, so there is a space after PATH: in my
filter. \S+ matches on any non-space characters. Nodes cannot have
spaces. It looks for the injection node (node1) to contain
paganini.bofh.team; however, the dot character in regex matches on any character, not just the period. To ensure a period character is in the
search, the dot is escaped with a backslash.
paganinin also has a Tor line into it, and why I optionally see if the
Tor side of paganini was the source. (!tor-network)? looks for 0 or 1
instance of "!tor-network". Sometimes injections nodes get demarked as
not for e-mail by adding "not-for-mail", so I check for the optional
substring that may appear zero or 1 times.
The $ anchors the substring to the right end of the string. You want to
test on node1, the injection node, not match on a different node. I
want to test when paganini is the source of an article, not if paganini
peered an article from elsewhere.
Path: ...!paganini.bofh.team
Path: ...!paganini.bofh.team!not-for-mail
Path: ...!paganini.bofh.team!tor-network!not-for-mail
Those are all possible injection node IDs for paganini. If Tbird does
not support regex to ensure you are testing on only the injection node
belongs to paganini, check if Tbird lets you test on custom headers that "contain a substring at the end of a head value". You could then test
on the 3 above substrings in the Path header.
You need to retrieve full messages. That means getting all headers and
body, not just the overview headers. You need to add a custom header
into Tbird. Path is not in the list of headers on which Tbird will
test. Tbird probably only tests on overview headers, and Path is a non-overview header. If Tbird doesn't support regex (unless you install
an add-on, like FiltaQuilla), you need to see if Tbird can test on a
substring at the *end* of a header's value.
https://support.mozilla.org/en-US/kb/organize-your-messages-using-filters
You need to add the Path custom header, so you can select it in the
Property field of a filter. Without regex support, you'd have to check
what tests can be selected in the Tests field. Without regex, you'd
need an "ends with" test. You would need 3 such filters to cover each
of the above possible injection node strings that paganini may insert
into the Path header. Only 1 filter would be needed if Tbird supported
regex.
https://www.betterbird.eu/support/
"Thunderbird does not recognise search terms containing regular
expression ..."
Presumably that statement means Betterbird supports regex. However, you probably should ask a Bbird user if it support regex. From:
https://www.betterbird.eu/releasenotes/index-previous-102.html
"Regular expressions in search terms (folder search and filters) via
"matches" and "doesn't match" conditions. Note that matching is case-insensitive and is implemented using the C++ std::regex library,
which (almost completely) follows the regular expression syntax in the
ECMA-262 Standard."
Sure looks like Bbird supports regex. Tbird does not unless you install
an add-on. I don't have Tbird to see if one of the tests you can select
is equivalent to "ends with" to let you test on the inject node in Path.
But that means you have to learn regex, and it is very complex. I've
been using regex in my filters for MANY years, and yet I've only scraped
the tip of the iceberg on what it can do.
The problem is using a combo client that does both e-mail and
newsgroups. They tend to focus on one (mostly e-mail), and the other
(NNTP) is an afterthought, and added only because the GUI layout for
e-mail facilitates the same layout usable for Usenet. A common GUI for
2 different protocols.
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)