Discussion:
Filter
(too old to reply)
FreeMan
2018-07-21 18:45:52 UTC
Permalink
Running Seamonkey for newsgroups.

How do I automatically filter crap (i.e. delete) from anon etc ?
VanguardLH
2018-07-21 18:58:30 UTC
Permalink
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
Can Seamonkey test on the PATH header? It's a non-overview header and
typically most NNTP clients only test on the overview headers.
FreeMan
2018-07-21 19:15:19 UTC
Permalink
Subject
From
Date
Size
Custom Header
There is NO local help for any of this !


Side note:
The developers have let us down on HELP and features that were extremely
important and provided by 3rd parties that now do not work !

And don't tell me about security. That is a false cognate.

In trying to protect use against ourselves is an endless taks so let
them screw up and learn.
Post by VanguardLH
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
Can Seamonkey test on the PATH header? It's a non-overview header and
typically most NNTP clients only test on the overview headers.
Ralph Fox
2018-07-21 21:32:00 UTC
Permalink
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:51.0) Gecko/20100101 Firefox/51.0 SeaMonkey/2.48
Subject
From
Date
Size
Custom Header
There is NO local help for any of this !
Here is the local help from my copy of SeaMonkey (version 2.49.3).
I found it between items 7 and 8 under "Creating Message Filters".

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

7. Use the drop-down lists to choose the search criteria (for example,
"Subject", "Sender", "contains", "doesn't contain") and then type
the text or phrase you want to match.

Tip: To search for messages that contain a header not listed in the
first drop-down menu (for example, if you want to search for
messages that include the header Resent-From), choose Customise and
type the header you want to search for. SeaMonkey Mail & Newsgroups
adds your custom header to the drop-down list, so you can then
choose it to search for matching entries. Make sure you enter the
custom header correctly, since Mail will only find entries that
exactly match what you type.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
--
Kind regards
Ralph
VanguardLH
2018-07-22 02:18:16 UTC
Permalink
Post by Ralph Fox
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:51.0) Gecko/20100101 Firefox/51.0 SeaMonkey/2.48
Subject
From
Date
Size
Custom Header
There is NO local help for any of this !
Here is the local help from my copy of SeaMonkey (version 2.49.3).
I found it between items 7 and 8 under "Creating Message Filters".
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
7. Use the drop-down lists to choose the search criteria (for example,
"Subject", "Sender", "contains", "doesn't contain") and then type
the text or phrase you want to match.
Tip: To search for messages that contain a header not listed in the
first drop-down menu (for example, if you want to search for
messages that include the header Resent-From), choose Customise and
type the header you want to search for. SeaMonkey Mail & Newsgroups
adds your custom header to the drop-down list, so you can then
choose it to search for matching entries. Make sure you enter the
custom header correctly, since Mail will only find entries that
exactly match what you type.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The valid syntax for a header is:

<headername>:<space><value>

Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.

For example, while I can test on non-overview headers in my client, Path
is not one of the standard headers in the client's list on which it will
check. So, to find someone posting through a remailer (anon), I use the
Header operator to specify the name of the header, as in:

Header {^Path: \S+(\bmail.?2.?news.*|\b(mixmin\.net|dizum\.net|reece\.net\.au|tioat\.net|mail2news\.nymu\.eu|gweep\.ca|bananasplit\.info|darner\.ixazon\.lan|tahina\.priv\.at|pseudonym\.nl)(!?\.?POSTED)?(!not-for-mail)?|!usenet-gateway|!sewer-output.*)$}

That's just one of my filters. That looks for a Path header (always
present and added by the server, NOT by the sender's client), headers
always start in column 1 (versus continuation lines that have a leading
space), followed by a colon and the one mandatory space. I skip a bunch
of the non-space characters for nodes before the last one in the list
which is actually the first node (insertion node) from where the post
originated. While this filter catches posts submitted through several
sources, like mixmin aka dizum, reece, and others, it gets too
complicated to write a regex to catch all the anonymizing sources, so I
have several rules to filter out the crap sources.

While it appears that Seamonkey can let you define a header name and
then use it in a filter, will Seamonkey actually test on non-overview
headers? The non-overview headers are not retrieved unless you use XPAT
or download the complete message (headers + body). XPAT is resource
intensive on the server (the client is asking the server to do a search)
so rare few NNTP servers support XPAT. In fact, I haven't come across
one yet in using multiple Usenet providers in over 20 years. Even my
client won't test on non-overview headers UNLESS it is configured to
download the entire message. I don't visit binary (porn) newsgroups, so
downloading all of every message in the text-only newsgroups doesn't
take very long, especially when just an incremental update (rather than
when first subscribing to a newsgroup that downloads all messages since
all will be new to the client). By configuring my client to download
the entire message, it can see the overview and non-overview headers and
also the body, so I could filter on any of that.

If you create a custom header name of Path and then try to test on that
header's value in Seamonkey, will it actually test on that non-overview
header?

When I use "telnet news.individual.net 119" to get a telnet session with
my NNTP server and then run "list overview.fmt", Path is not listed as
an overview header. I doubt it is considered an overview header at
other NNTP servers. Because XPAT isn't supported anywhere (client might
have it but the servers don't support it or honor such requests), I've
had to download the complete message to get at the non-overview headers.
VanguardLH
2018-07-22 02:23:09 UTC
Permalink
Post by VanguardLH
Post by Ralph Fox
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:51.0) Gecko/20100101 Firefox/51.0 SeaMonkey/2.48
Subject
From
Date
Size
Custom Header
There is NO local help for any of this !
Here is the local help from my copy of SeaMonkey (version 2.49.3).
I found it between items 7 and 8 under "Creating Message Filters".
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
7. Use the drop-down lists to choose the search criteria (for example,
"Subject", "Sender", "contains", "doesn't contain") and then type
the text or phrase you want to match.
Tip: To search for messages that contain a header not listed in the
first drop-down menu (for example, if you want to search for
messages that include the header Resent-From), choose Customise and
type the header you want to search for. SeaMonkey Mail & Newsgroups
adds your custom header to the drop-down list, so you can then
choose it to search for matching entries. Make sure you enter the
custom header correctly, since Mail will only find entries that
exactly match what you type.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ QUOTE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<headername>:<space><value>
Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.
For example, while I can test on non-overview headers in my client, Path
is not one of the standard headers in the client's list on which it will
check. So, to find someone posting through a remailer (anon), I use the
Header {^Path: \S+(\bmail.?2.?news.*|\b(mixmin\.net|dizum\.net|reece\.net\.au|tioat\.net|mail2news\.nymu\.eu|gweep\.ca|bananasplit\.info|darner\.ixazon\.lan|tahina\.priv\.at|pseudonym\.nl)(!?\.?POSTED)?(!not-for-mail)?|!usenet-gateway|!sewer-output.*)$}
That's just one of my filters. That looks for a Path header (always
present and added by the server, NOT by the sender's client), headers
always start in column 1 (versus continuation lines that have a leading
space), followed by a colon and the one mandatory space. I skip a bunch
of the non-space characters for nodes before the last one in the list
which is actually the first node (insertion node) from where the post
originated. While this filter catches posts submitted through several
sources, like mixmin aka dizum, reece, and others, it gets too
complicated to write a regex to catch all the anonymizing sources, so I
have several rules to filter out the crap sources.
While it appears that Seamonkey can let you define a header name and
then use it in a filter, will Seamonkey actually test on non-overview
headers? The non-overview headers are not retrieved unless you use XPAT
or download the complete message (headers + body). XPAT is resource
intensive on the server (the client is asking the server to do a search)
so rare few NNTP servers support XPAT. In fact, I haven't come across
one yet in using multiple Usenet providers in over 20 years. Even my
client won't test on non-overview headers UNLESS it is configured to
download the entire message. I don't visit binary (porn) newsgroups, so
downloading all of every message in the text-only newsgroups doesn't
take very long, especially when just an incremental update (rather than
when first subscribing to a newsgroup that downloads all messages since
all will be new to the client). By configuring my client to download
the entire message, it can see the overview and non-overview headers and
also the body, so I could filter on any of that.
If you create a custom header name of Path and then try to test on that
header's value in Seamonkey, will it actually test on that non-overview
header?
When I use "telnet news.individual.net 119" to get a telnet session with
my NNTP server and then run "list overview.fmt", Path is not listed as
an overview header. I doubt it is considered an overview header at
other NNTP servers. Because XPAT isn't supported anywhere (client might
have it but the servers don't support it or honor such requests), I've
had to download the complete message to get at the non-overview headers.
Oops, that should be the XHDR command (from the client), not the XPAT
command. XPAT would be pattern search at the server to find a message.
XHDR would be used with a specific message ID to get a specific header
from that message. As with XPAT, I've not happened to use an NNTP
server that support XHDR. Those are server-side resource intensive
operations. Most Usenet providers don't want their users consuming all
those resources, plus getting at all the headers or doing pattern
searching is something the client can (or should) do.
Ralph Fox
2018-07-22 05:20:13 UTC
Permalink
Post by VanguardLH
<headername>:<space><value>
To: FreeMan

To avoid confusion, the above does NOT apply to SeaMonkey filters.

Here is a screen-shot of setting up a kill filter in SeaMonkey
to automatically delete posts from anon:

<Loading Image...>

* The header name is in column 1, *without* a colon after it.
* The value is in column 3, without a space before it.
* Column 2 has the test to perform - in this case, "contains".


The <headername>:<space><value> applies to when you look at the
headers in the message source in SeaMonkey
View >> Message Source (Ctrl+U)
Post by VanguardLH
Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.
To: FreeMan

To avoid confusion, SeaMonkey itself does not support regex.
SeaMonkey also does not need regex to match on a substring.
--
Kind regards
Ralph
VanguardLH
2018-07-23 03:25:19 UTC
Permalink
Post by Ralph Fox
Post by VanguardLH
<headername>:<space><value>
To: FreeMan
To avoid confusion, the above does NOT apply to SeaMonkey filters.
Here is a screen-shot of setting up a kill filter in SeaMonkey
<https://i.imgur.com/IHbcbAX.png>
* The header name is in column 1, *without* a colon after it.
* The value is in column 3, without a space before it.
* Column 2 has the test to perform - in this case, "contains".
The <headername>:<space><value> applies to when you look at the
headers in the message source in SeaMonkey
View >> Message Source (Ctrl+U)
Post by VanguardLH
Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.
To: FreeMan
To avoid confusion, SeaMonkey itself does not support regex.
SeaMonkey also does not need regex to match on a substring.
The problem then is that, say, filtering on "mixman" will fire on any
substring in Path with "mixman" anywhere within it. You want to filter
out posts that *originate* from a mixman server, not posts that peered
through mixman servers. Someone could submit to, say, individual.net
who peers to many NNTP servers who also peer to others. When you
retrieve articles from your NNTP server, one of the peering routes
might've gone through a mixman server - but it did not *originate* from
there! I filter on anon posts submitted to mixman servers, not peered
through them and that would result in a lot false positives.

Without bounding the "mixman" string by delimiters present in the Path
header and without fixing its position at the end of the Path header to
identify it as the injection node, the OP with Seamonkey would be
filtering out posts that were never submitted to a mixman server but
merely peered through one.

Pity Seamonkey doesn't support regex. It allows focusing on WHERE
within a string that a substring will appear for firing a rule.
Ralph Fox
2018-07-23 08:15:21 UTC
Permalink
Post by VanguardLH
Post by Ralph Fox
Post by VanguardLH
<headername>:<space><value>
To: FreeMan
To avoid confusion, the above does NOT apply to SeaMonkey filters.
Here is a screen-shot of setting up a kill filter in SeaMonkey
<https://i.imgur.com/IHbcbAX.png>
* The header name is in column 1, *without* a colon after it.
* The value is in column 3, without a space before it.
* Column 2 has the test to perform - in this case, "contains".
The <headername>:<space><value> applies to when you look at the
headers in the message source in SeaMonkey
View >> Message Source (Ctrl+U)
Post by VanguardLH
Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.
To: FreeMan
To avoid confusion, SeaMonkey itself does not support regex.
SeaMonkey also does not need regex to match on a substring.
The problem then is that, say, filtering on "mixman" will fire on any
substring in Path with "mixman" anywhere within it.
Not so.
In SeaMonkey the match criteria can be something other than "contains".

For example the match criteria can be "ends with", as in this screen-shot
of a SeaMonkey filter:

<Loading Image...>
Post by VanguardLH
You want to filter
out posts that *originate* from a mixman server, not posts that peered
through mixman servers. Someone could submit to, say, individual.net
who peers to many NNTP servers who also peer to others. When you
retrieve articles from your NNTP server, one of the peering routes
might've gone through a mixman server - but it did not *originate* from
there! I filter on anon posts submitted to mixman servers, not peered
through them and that would result in a lot false positives.
Without bounding the "mixman" string by delimiters present in the Path
header and without fixing its position at the end of the Path header to
identify it as the injection node, the OP with Seamonkey would be
filtering out posts that were never submitted to a mixman server but
merely peered through one.
Pity Seamonkey doesn't support regex. It allows focusing on WHERE
within a string that a substring will appear for firing a rule.
One does not require regex to do this. For example, in SeaMonkey one
can create a filter like this:

Header: Path
Match criteria: ends with
Value: !news.mixman.net!.POSTED!not-for-mail
--
Kind regards
Ralph
VanguardLH
2018-07-23 18:54:15 UTC
Permalink
Post by Ralph Fox
Post by VanguardLH
Post by Ralph Fox
Post by VanguardLH
<headername>:<space><value>
To: FreeMan
To avoid confusion, the above does NOT apply to SeaMonkey filters.
Here is a screen-shot of setting up a kill filter in SeaMonkey
<https://i.imgur.com/IHbcbAX.png>
* The header name is in column 1, *without* a colon after it.
* The value is in column 3, without a space before it.
* Column 2 has the test to perform - in this case, "contains".
The <headername>:<space><value> applies to when you look at the
headers in the message source in SeaMonkey
View >> Message Source (Ctrl+U)
Post by VanguardLH
Some clients will let you specify just the <headername>, like Path.
However, if you're not matching on a standard header (overview or
non-overview) then you might have to use regex to match on a substring,
like ^<headername>:\s where ^ anchors to column 1 (start of line), you
add the colon character, and specify the intervening space between
header name and its value.
To: FreeMan
To avoid confusion, SeaMonkey itself does not support regex.
SeaMonkey also does not need regex to match on a substring.
The problem then is that, say, filtering on "mixman" will fire on any
substring in Path with "mixman" anywhere within it.
Not so.
In SeaMonkey the match criteria can be something other than "contains".
For example the match criteria can be "ends with", as in this screen-shot
<https://s8.postimg.cc/b1923lwdx/Sea_Monkey-filter-to-match-end-of-_Path-header.png>
Post by VanguardLH
You want to filter
out posts that *originate* from a mixman server, not posts that peered
through mixman servers. Someone could submit to, say, individual.net
who peers to many NNTP servers who also peer to others. When you
retrieve articles from your NNTP server, one of the peering routes
might've gone through a mixman server - but it did not *originate* from
there! I filter on anon posts submitted to mixman servers, not peered
through them and that would result in a lot false positives.
Without bounding the "mixman" string by delimiters present in the Path
header and without fixing its position at the end of the Path header to
identify it as the injection node, the OP with Seamonkey would be
filtering out posts that were never submitted to a mixman server but
merely peered through one.
Pity Seamonkey doesn't support regex. It allows focusing on WHERE
within a string that a substring will appear for firing a rule.
One does not require regex to do this. For example, in SeaMonkey one
Header: Path
Match criteria: ends with
Value: !news.mixman.net!.POSTED!not-for-mail
That works to test if mixmin is the injection node, not a peer.

Ralph Fox
2018-07-21 22:03:43 UTC
Permalink
Post by VanguardLH
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
Can Seamonkey test on the PATH header? It's a non-overview header and
typically most NNTP clients only test on the overview headers.
Seamonkey can test on the PATH header provided the news server supports
the XHDR command for non-overview headers such as the PATH header. Some
news servers do, some don't.

Today, the big "premium" news servers (which cater for binary users)
generally do NOT support the XHDR command for non-overview headers.

I don't know whether or not the OP's news server freenews.netfront.net
(which is based in Hong Kong) supports the XHDR command for non-overview
headers.
--
Kind regards
Ralph
Johnny
2018-07-21 19:11:40 UTC
Permalink
On Sat, 21 Jul 2018 11:45:52 -0700
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
Get a good newsreader that will filter on a phrase in the headers, and
take the time to learn how to use it.

Filter dizum.com and mixmin.net and that will get rid of most of the
garbage.

https://www.claws-mail.org/downloads.php?section=downloads
G Ross
2018-07-21 19:15:09 UTC
Permalink
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
I could not find a way to delete such, but I filter on the name and
mark it as read.
--
G Ross
R.Wieser
2018-07-21 19:46:09 UTC
Permalink
FreeMan,
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
As you have not mentioned which version you are using the below links might
not be exactly what you are looking for, but maybe they gives enough clues
to what you want.

http://www.dwx.com/support/faqs/seamonkey/seamonkeyrules.php

http://www.freeemailtutorials.com/mozillaMail/spamJunkMail.cwd

I found the above on the first result page googeling for "Seamonkey blocking
sender" (without the quotes)

Regards,
Rudy Wieser
Ralph Fox
2018-07-21 22:11:45 UTC
Permalink
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:51.0) Gecko/20100101 Firefox/51.0
SeaMonkey/2.48
Latest Seamonkey.
Not according to your headers.

Latest SeaMonkey is version 2.49.3.
Your headers say you are running version 2.48.
--
Kind regards
Ralph
Daniel60
2018-07-22 08:25:19 UTC
Permalink
Post by FreeMan
Running Seamonkey for newsgroups.
How do I automatically filter crap (i.e. delete) from anon etc ?
Freeman, although there has been some very useful suggestions made here,
if you were to set yourself up a news account to news.mozilla.com on
Port 119 and then subscribe to the mozilla.support.seamonkey news group,
you might also get some informative answers from other SeaMonkey users
... and developers, as well.

I believe that groups.google.com also carries the
mozilla.support.seamonkey news group.

And, as mentioned by others, the latest release version of SeaMonkey is
2.49.3 .... and 2.49.4 could be out any day now.
--
Daniel
Loading...