🚀 go-pugleaf

On 9 Feb, in article
     <PM000564C...@aca41494.ipt.aol.com> See....@aol.com
     "jmfbahciv" wrote:

> Andy Burns wrote:
>> jmfbahciv wrote:
>>
>>> Niklas Karlsson wrote:
>>>
>>>> Your Name Here wrote:
>>>>
>>>>> This is a horrible computer ….
>>>
>>> I did not type that weird character.
>>
>> It seems you (or your slippery new computer did)
>>
>> The sequence "…." without quote marks appears in your
>> original message which may be a garbled unicode sequence for a
>> three-dotted ellipse, followed by a fullstop
>>
>> (U+2026 = 0xE2 0x80 0xA6) (0x2E)
>>
>> see 6th row on left here
>> <http://www.i18nqa.com/debug/utf8-debug.html>
>>
>> So presumably you typed "...." and something not quite clever enough for
>> its own good altered it?  Niklas's quote didn't mangle your original
>> characters, but your re-quoting of Niklas's reply changed it to
>> "…."
>>
>> So I think you have some kind of 8-bit character set/encoding problem
>> going on ...
> 
> Aw, fuckmeverymuch.  Thanks for the info.  Another item on the list of
> things to undo.
> 
> /BAH

Water way under the bridge by now, and noting that BAH is currently
off-line anyway, but just for info: the fundamental cause of BAH's 8-bit
character set problem is Niklas's non-compliant slrn/pre0.9.9-111, which
incorrectly used "charset=utf8" in the MIME Content-Type field: the
correct charset parameter for UTF-8 has a hyphen (RFC 3629 section 8).

The only problem with BAH's setup is that her Pineapple News/0.9.0
didn't recognise the invalid "utf8" charset, and hence it displayed a
three-byte UTF-8 encoding in Niklas's malformed message as three
separate characters, and encoded them thus when sending her reply.

This seems to be caused by a faulty example in old slrn.rc documentation
rather than a bug in slrn per se, so Niklas should be able to fix his
config; there are references to the docs being fixed in slrn in 2010,
and a 2011 Debian bug report of the same issue notes that

  a wrong charset definition like that will keep other newsreaders from
  correctly displaying the message (as they don't know a charset named
  "utf8" and fall back to their default which very likely is not UTF-8).

which is exactly what we've seen here.

You can certainly argue that Pineapple News could be more accepting of
malformed charsets (as some others -- including slrn -- evidently are),
but it's a little harsh, and sadly misleading, to tell BAH that she's
got an 8-bit charset/encoding problem -- or to suggest that her Mac
isn't Unicode-compliant, as someone else did -- when the fundamental
fault lies elsewhere.

(Hope you get to see this somehow, BAH!)

-- 
Simon Turner                                    DoD #0461
si...@twoplaces.co.uk
    Trust me -- I know what I'm doing!          -- Sledge Hammer