[lojban-beginners] Re: more 'suck'

[der Mouse <mouse@Rodents.Montreal.QC.CA>]

>> Speaking of which, could someone possibly convince the list to stop
>> emitting mail with headers containing non-ASCII characters?
> No, because the list doesn't generate the messages, it just passes
> them on with as little munging as possible.

Then I'd say it shouldn't be accepting them, if that's what it takes to
stop it from emitting them.

> Furthermore, how would you convert them?

Convert them?  I wouldn't.  I'd bounce them.

Indeed, I *do* bounce them.

Check the list logs and you'll probably see rejections, my mailer
saying "554 Invalid character in header", for messages containing

	From: Jorge "Llambías" <jjllambias2000@yahoo.com.ar>

(where I've simply copied the octets from the logged header to this
message, causing them to be taken as 8859-1 because that's how I've
marked the message).  (Based on message I see quoting other messages I
didn't see, I assume this is xorxes, this reinforced by the close match
between "xorxes" and "Jorge".)

> Just dropping all the non-ascii would change the meaning quite a bit,
> and I'm not sure how any other system would work.

I suspect encoding them assuming 8859-1 would work better, for purely
pragmatic values of "better".  Checking to see if the message itself
has a character set declared and, if so using that, would probably work
"better" yet.

I prefer bouncing them, though; contriving your systems to paper over
third parties' brokennesses only piles up trouble for the future.
Render their brokennesses *blatantly* broken and they're much more
likely to get fixed.

> Anyways, ecartis seems to have no relevant options.  If you have an
> idea of how a list *should* behave in this case, I can ask the
> developers to take a look at it.

My idea of how a list - or any other mailer - should behave when handed
such a message is simple: it should hard-reject it (as mine does).  The
only other even vaguely reasonable things I can see to do are to drop
the invalid characters, MIME-encode them assuming some character set
(vide supra), or convert them to something like question marks, and I
don't really like any of those.  I suppose you could drop the offending
header altogether, but that is even less likely to produce sensible
results.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B