🚀 go-pugleaf

In article  <20070929.1509.70906204snz$@i-m-t.demon.co.uk>
           v$af$ppint@i-m-t.demon.co.uk "ppint. at IMT" writes:

>      - hi; in dipsp article, <1190697209snz@nospam.demon.co.uk>,
>              pete@nospam.demon.co.uk observed & enquired:
> >        v$af$ppint@i-m-t.demon.co.uk "ppint. at IMT" writes:
> >>       pete@nospam.demon.co.uk up and announced (in parts):
> >>>
> >[snip utf-8 stuff]
> >>>Off to investigate Unicode now...
> >>>
> >>      - that took me a bit by surprise, that did - and then it
> >>      wasn't possible to connect to demon by dial-up a] reliably,
> >>      b] on 08440416662, c] on 08440996662, and/or any other way
> >>      from 015395-624xx and/or 01524-3821xx for...
> >>                       [it still isn't :-((]
> >
> >I trust it's better now?  Since I came back from hols (which is 
> >my excuse for the tardy response) things here seem to be back to 
> >"normal" re: authentication and routing.

How suddenly things change...  I haven't been able to get through 
to any number I've tried between 0800 and 1800 (from 0118979xxxx). 
Methinks I'll take a visit to www.demon on the XP box and check 
that the SDU numbers haven't been changed while I wasn't looking.

>         - ummm; from 01524-3821xx i can connect via 08440416662 to
>         anchor-du-22, 23 or 24 (iirc) approximately one dial-up in
>         two; the other attempts give what i believe to be the sig-
>         nal for "engaged" or "line busy"; this is not entirely sat-
>         isfactory, but is a *vast* improvement;
> 
>         - from 015395-624xx i cannot of late reliably comment:
>         i-m-t-at-home, which has, for my sins, been a windows 2000
>         professional (if that's not an oxymoronic cognomen) machine
>         (to allow me to make use of paypal, livejournal, & the like)
>         after complaining about being asked to perform its various
>         accustomed functions - increasingly slowly for a while - has
>         ceased to function at all; and i do not have another, fully
>         set-up box working at home (i should have another, strictly
>         dos 486 box set up tonight/tomorrow, some time; but whilst
>         this should see me back online from home reliably enough, in-
>         definitely, it'll not solve the minor matters of lj, paypal,
>         and other more complicated sites' demands upon browsers, that
>         gpl arachne cannot [- yet? -] handle.)

Think I read somewhere recently that there was a newer version of 
Arachne available.  Not that is will necessarily solve the https: 
issues, but even if it did most websites these days are not usuable 
over a narrowband dialup.

[..]
[utf-8 stuff]
> >>      - um, any update on the ease (or otherwise) of tackling
> >>      this, and possibly even upon your having accomplished it,
> >>      awaited with interest (but i cannot promise bated breath).
> >
> >Still "work in progress"[*].  I might get to take a look this w/e 
> >and see how simply ignoring the top 8 bits of the Unicode char 
> >works (or most likely doesn't...)
> >
> >[*] which of course means I haven't even started yet :-(
> 
>         - whilst it's desirable, and would certainly improve the
>         elegance of the presentation, as well as the functionality,
>         it's hardly a matter, over delays to which you should re-
>         proach yourself: an occasional dribblet of news, or else
>         the periodic "nothing yet" from time to time... ;-)

Well.  There is a teensy-weensy problem: the Unicode characters 
were (I thought) 16-bit, but it appears that they are getting 
wider than that.  But no matter; even 64K characters are beyond 
the capability of DOS -- where we have only 256 at a time :-(

However, all is not lost.  I've thrown together some ugly^H^H^H^Helegant code
which pretends that "charset=UTF-8" is "charset=ISO-8859-1" 
(which is I guessed what most of us Eurotypes are interested in) 
and mapped the common Unicode punctuation code points (U+2000 to 
U+206F) to the nearest ASCII displayable equivalent.  Everything 
else gets mapped to an inverted question mark ('�' or U+00BF).  
Seems to work, where "work" is relative rather than absolute :)  
At least it renders the article slightly more readable, and 
consumes only a few hundred bytes of memory.  If there is anyone 
maintaining the JLW Snews (are you listening, John?) who would 
like to see my code with a view to implementing it, just let me 
know. [My Snews is based on an earlier GeK version and is so 
different from the JLW one that this task might not be as 
straightforward as a copy&paste job...]

> >>      - your next task, "should you choose to accept it", is to
> >>      identify the simplest way of allowing the use of regexps
> >>      in the news kill file... *g*
> >
> >Which version of NET are you using?  I'm using the latest 
> >official GremNOS here, and notice that my news kill file contains 
> >regexps -- well, the simplified regexps that wildmat() handles 
> >anyway like '*', '?' and '[...]' sets.  Seems to work for me.
> >
>         - i am, i believe, using pplt_03d.zip (unzipped), kindly
>         crafted for me because some lowlife's taken to targeting
>         postmaster at my d[e,o]m[o,(ai)]n...

Ah yes -- I remember now.

>                - occasionally hundreds of time a day.
> 
>         - it certainly gives the impression of not understanding
>         regexps to the extent that, by contrast, work in mailkill.
>           most recently, i've been a little dismayed to discover
>         that trolls & spammers placing an extra space in between
>         the From: and the overt poster is sufficient for them to
>         evade all my news killfile rules testing this line. :-(

AFAICR the main restriction is that news killfile entries cannot 
start with a regexp -- the first character must be 'a' to 'z' 
(not sensitive to case).  FWIW here are a few examples from my 
killfile:

  # kill crossposts to 4 or more groups
  Newsgroups:*,*,*,*
  # kill any post that includes uk.misc*
  Newsgroups:*uk.misc*
  # keep dispc unconditionally
  !Newsgroups:*demon.ip.support.pc*
  !Newsgroups:*comp.risks*
  Subject:*!!!!!*
  Content-type:*html*
  # kill any post 400 lines or more
  Lines:*[4-9][0-9][0-9]*
  Lines:*[0-9][0-9][0-9][0-9]*

Note the lack of <space> after the colon; spaces or tabs will be 
gobbled up with the '*'.  Note too that "keep" [!] rules take 
precedence over "kill" rules (thus ensuring that I still receive 
comp.risks postings) and that the order in which keep/kill rules 
are specified does not matter.

HTH,
Pete
-- 
   demon.ip.support.pc is for discussion of IP-related issues
   using plain DOS software - See the Which Group FAQ:-
   ftp://ftp.demon.co.uk/pub/doc/general/whichgrp.faq