[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [leafnode-list] Why was support for 8 bit characters in headers



"Peter N. M. Hansteen" <peter@xxxxxxxxx> writes:

> In the dk.* and no.* regional hierarchies at least, posting those munged
> headers is the surest path available to killfiles and flame wars. The

Oh, and omitting RFC-compliant encoding is going to get you nasty
bashing in de.*, so there ;-)

> reason is simple: All news readers will render something more or less
> readable, ie "Bjørnen sover, men snorker fælt" might at worst be
> rendered as "Bj|rnen sover, men snorker f[lt". The result is at least
> readable. Not so if the reader does not implement RFC2047 properly -
> AFAIK most of them.

Most current Unix newsreaders support RFC-2047, and the big graphical
clients (Outlook, Mozilla) do, more or less (modulo Microsoft bugs, of
course).

It's not leafnode's fault if your favourite newsreader does not support
that -- and IF the national characters really was that big a problem,
you'd long gone for æ -> ae, ø -> oe, å -> aa (that's for DK at best I
can tell, knowing only tiny fragments of bokmål and dansk, enough to
read simple signs basically), which would certainly work out nicely in
the header. Been there, done that with German some years ago (ä -> ae, ö
-> oe, ü -> ue, ß -> ss (sz in Switzerland and by some clueless Germans
and Austrians)).

RFC-2047 is the way to transport that stuff, even across replies, those
their readers support RFC-2047 will see "Re: Bjørnen sover, men snorker
fælt" even if the reader of the person who sent the followup had
absolutely no clue. Just slapping "Re: " and taking the rest of the
header unchanged is fine.

>> The very moment UTF-8 will be ratified as official RFC (it's in the
>> development process and an Internet Draft of the USEFOR group), we will
>> allow 8-bit characters again. Feel free to bug me should I miss the
>> future release of the RFC that supersedes RFC-1036.
>
> You seem to be under the impression that UTF-8 is close to getting
> accepted. I for one think this is naive at best.

Well, sending anything other than US-ASCII or UTF-8 out in headers is
plain broken. I may add an option to let garbage out, with a warning
attached to the header, similar to the X-Authentication-Warning that
sendmail uses under certain circumstances.

> And in the meantime, the problem is that in certain regional
> hierarchies,
>
>   We! Need! 8bit! support! Now!

The more bangs, the easier ignored ;-)

> Yes, there is quite a significant amounts of work to be done before all
> those issues are resolved and implemented as anything approaching sane
> solutions.
>
> Throwing out 8-bit headers now in all probability means waiting several
> years. If an UTF-8 based solution does indeed materialize, that is.

Well, newsreaders that ignore that some users and many other newsreaders
have and support national characters by means of RFC-2045..2049, some
even RFC-2231 (which I consider wrong, but it's there), will have to
suffer sooner or later and should be phased out and replaced by more
capable software. (In the de.* local Usenet hierarchy, 8bit is the
recommended encoding, much preferred over quoted-printable, not only
because Outlook hoses quoting of base64/qp.)

> I reiterate - dropping support for 8 bit headers was premature.

Sending 8bit headers without proper declaration was premature, at
best. I am seriously considering the option mentioned above, to 1.9.25,
call it "hey leafnode close your eyes even if the user sends garbage"
and have it disclaim responsibility if garbage is sent out.

I go for RFC-2045/2047.

-- 
Matthias Andree

-- 
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list