MT-NW Regular Expression Filtering

Dave Allen wrote:

> I'm not great with regular expressions. Is there a regular expression I
> can use with MT-NewsWatcher to match these hashed up subject lines?

The regular expression "M.*I.*5" should do it, but it would also catch a
subject line like "Let's hope there's no 'Mission: Impossible' 4 or 5".
"M.?I.?5" will catch the current variations, but won't work if the
spammer puts more than one character between the M, the I, or the 5.
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article ,
Michelle Steiner wrote:

> In article ,
> Dave Allen wrote:
>
> > I'm not great with regular expressions. Is there a regular
> > expression I can use with MT-NewsWatcher to match these hashed up
> > subject lines?
>
> I haven't used regex at all, but after studying some documentation on it
> for the past few minutes, I wonder whether this will do it:
>
> M[^]I[^]5

I think you misunderstood something. The expression [^] is not valid
syntax. If it were it would mean all characters not matching the empty
set.

--
Tom Stiller

PGP fingerprint = 5108 DDB2 9761 EDE5 E7E3 7BDA 71ED 6496 99C0 C7CF
 >> Stay informed about: MT-NW Regular Expression Filtering 

This message is not archived
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article
,
Dave Allen wrote:

> The MI5 idiot used to be pretty easy to filter out. Nearly all of his
> posts could be killed by matching simple filters like:
>
> "from" contains "MI5Victim@mi5.gov.uk"
> "subject" contains "MI5 Persecution"
>
> The subject filter also caught replies to his posts.
>
> Unfortunately, his more recent posts all have different "from" addresses
> and his latest subject lines are all hashed up.
>
> I'm not great with regular expressions. Is there a regular expression I
> can use with MT-NewsWatcher to match these hashed up subject lines?
>
> Subject: M'I.5'P ersecution BBC Newsca sters L ie & De ny T heyre
> Watc hing Me
>
> Subject: M-I'5.Persecuti on . BBC Newscasters L ie & D eny Theyre
> Watchin g Me
>
> Subject: M,I-5'Persecutio n , B BC New scasters Li e & Deny T heyre Wa
> tching Me
>
> Subject: M-I,5.Perse cution - Molestat ion dur ing Trav el
>
> Subject: M`I'5-Persecuti on . Mole station dur ing Tr avel
>
> Subject: M-I 5.Perse cution . Four Ye ars of MI5 Persecuti on Po sts
> on I nternet Newsgr oups

This seems to be working

M[^a-z0-9]*I[^a-z0-9]*5[^a-z0-9]*P[^a-z0-9]*e[^a-z0-9]*r[^a-z0-9]*s

It is looking for MI5Pers with zero or more anythings that are not
a-z A-Z or 0-5 between each letter.

I guess I choose to be less aggressive and included the beginning
of Persecution "Pers". If you are satisfied that MI5 is
sufficient, then drop the stuff from the 'P' on.

Bob Harris
 >> Stay informed about: MT-NW Regular Expression Filtering 

This message is not archived
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article ,
Bob Harris wrote:

> This seems to be working
>
> M[^a-z0-9]*I[^a-z0-9]*5[^a-z0-9]*P[^a-z0-9]*e[^a-z0-9]*r[^a-z0-9]*s

I don't think he'd use anything but punctuation that lets his subject
line be reasonably easy to read, so I use this which seems to work:

..*M[~`'" -_,]I[~`'" -_,]5.*

There's a space in the character class. Note that I don't care _what_
comes after the 5. You may be able to get away with [:punct:]; I'm not
sure it works in MTNW.
--
W. Oates
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article ,
Michelle Steiner wrote:

> In article ,
> Tom Stiller wrote:
>
> > > I haven't used regex at all, but after studying some documentation
> > > on it for the past few minutes, I wonder whether this will do it:
> > >
> > > M[^]I[^]5
> >
> > I think you misunderstood something. The expression [^] is not valid
> > syntax. If it were it would mean all characters not matching the
> > empty set.
>
> That's what I thought it meant. If that were a valid construct, it
> would catch everything containing MI5, with anything between the
> letters, wouldn't it?

I think the syntax you meant is:
M[^I]*I[^5]*5

which is an 'M', followed by zero or more characters other than an 'I',
followed by an 'I', followed by zero or more characters other than a
'5', followed by a '5'.
>
> BTW, as soon as I read "sets" in the article, things began to make sense
> to me because I know a bit about set theory. Or at least, I think I do.

--
Tom Stiller

PGP fingerprint = 5108 DDB2 9761 EDE5 E7E3 7BDA 71ED 6496 99C0 C7CF
 >> Stay informed about: MT-NW Regular Expression Filtering 

This message is not archived
 >> Stay informed about: MT-NW Regular Expression Filtering 

Dave Allen wrote:

> The MI5 idiot used to be pretty easy to filter out. Nearly all of his
> posts could be killed by matching simple filters like:
>
> "from" contains "MI5Victim@mi5.gov.uk"
> "subject" contains "MI5 Persecution"
>
> The subject filter also caught replies to his posts.
>
> Unfortunately, his more recent posts all have different "from" addresses
> and his latest subject lines are all hashed up.
>

snip
the Berlin University Usenet news server filters him out

I never heard of him until I saw him in google groups

Hugh W


--
For genealogy and help with family and local history in Bristol and
district http://groups.yahoo.com/group/Brycgstow/

http://snaps4.blogspot.com/ photographs and walks

GENEALOGE http://hughw36.blogspot.com/ MAIN BLOG
 >> Stay informed about: MT-NW Regular Expression Filtering 

This message is not archived
 >> Stay informed about: MT-NW Regular Expression Filtering 

Warren Oates wrote:
> In article ,
> Bob Harris wrote:
>
>> This seems to be working
>>
>> M[^a-z0-9]*I[^a-z0-9]*5[^a-z0-9]*P[^a-z0-9]*e[^a-z0-9]*r[^a-z0-9]*s
>
> I don't think he'd use anything but punctuation that lets his subject
> line be reasonably easy to read, so I use this which seems to work:
>
> .*M[~`'" -_,]I[~`'" -_,]5.*

FWIW, I just developed something similar for tin using the wildmat
syntax. Here's how it appears in my ~/.tin/filter filter:

comment=MI5 wildcard
group=*
case=0
score=kill
subj=*M[-`'., ]I[-`',. ]5[-`',. ]*

Note that the placement of the dash first in the set is significant
(ie. it means something else if it is not first) and that I haven't
included a few of the characters that Warren has, but it would be
trivial to do so.

--
*--------------------------------------------------------*
| ^Nothing is foolproof to a sufficiently talented fool^ |
| Heath Raftery, HRSoftWorks _\|/_ |
*______________________________________m_('.')_m_________*
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article ,
Heath Raftery wrote:

> Warren Oates wrote:
> > In article ,
> > Bob Harris wrote:
> >
> >> This seems to be working
> >>
> >> M[^a-z0-9]*I[^a-z0-9]*5[^a-z0-9]*P[^a-z0-9]*e[^a-z0-9]*r[^a-z0-9]*s
> >
> > I don't think he'd use anything but punctuation that lets his subject
> > line be reasonably easy to read, so I use this which seems to work:
> >
> > .*M[~`'" -_,]I[~`'" -_,]5.*
>
> FWIW, I just developed something similar for tin using the wildmat
> syntax. Here's how it appears in my ~/.tin/filter filter:
>
> comment=MI5 wildcard
> group=*
> case=0
> score=kill
> subj=*M[-`'., ]I[-`',. ]5[-`',. ]*
>
> Note that the placement of the dash first in the set is significant
> (ie. it means something else if it is not first) and that I haven't
> included a few of the characters that Warren has, but it would be
> trivial to do so.


Back when this [MI5] became a problem in december i crafted this:
kill if subj= m.i.5.p (case insensitive)

and it worked pretty well. I tested it on about 10k newsgroup articles
and didn't get any false hits (athough they are certainly possible with
this filter) nor any misses.

I chose this route so i wouldn't have to find all the non-alpha
characters on my keyboard. :-) That, and i don't really know how to use
regexps.

..max

--
The part of betatron @ earthlink . net was played by a garden gnome
 >> Stay informed about: MT-NW Regular Expression Filtering 

In article
,
max wrote:

> In article ,
> Heath Raftery wrote:
>
> > Warren Oates wrote:
> > > In article ,
> > > Bob Harris wrote:
> > >
> > >> This seems to be working
> > >>
> > >> M[^a-z0-9]*I[^a-z0-9]*5[^a-z0-9]*P[^a-z0-9]*e[^a-z0-9]*r[^a-z0-9]*s
> > >
> > > I don't think he'd use anything but punctuation that lets his subject
> > > line be reasonably easy to read, so I use this which seems to work:
> > >
> > > .*M[~`'" -_,]I[~`'" -_,]5.*
> >
> > FWIW, I just developed something similar for tin using the wildmat
> > syntax. Here's how it appears in my ~/.tin/filter filter:
> >
> > comment=MI5 wildcard
> > group=*
> > case=0
> > score=kill
> > subj=*M[-`'., ]I[-`',. ]5[-`',. ]*
> >
> > Note that the placement of the dash first in the set is significant
> > (ie. it means something else if it is not first) and that I haven't
> > included a few of the characters that Warren has, but it would be
> > trivial to do so.
>
>
> Back when this [MI5] became a problem in december i crafted this:
> kill if subj= m.i.5.p (case insensitive)
>
> and it worked pretty well. I tested it on about 10k newsgroup articles
> and didn't get any false hits (athough they are certainly possible with
> this filter) nor any misses.
>
> I chose this route so i wouldn't have to find all the non-alpha
> characters on my keyboard. :-) That, and i don't really know how to use
> regexps.
>
Try

M[^I]*I[^5]*5

which is an 'M', followed by zero or more characters other than an 'I',
followed by an 'I', followed by zero or more characters other than a
'5', followed by a '5'.

--
Tom Stiller

PGP fingerprint = 5108 DDB2 9761 EDE5 E7E3 7BDA 71ED 6496 99C0 C7CF
 >> Stay informed about: MT-NW Regular Expression Filtering