Re: MIME encoding of headers

In message <[email protected]>, Valdis.Kletni
[email protected] writes:
>On Sun, 23 May 2004 18:44:13 +0200, Christer Borang <[email protected]>  s
>aid:
>> I would like to tell exmh to use 8bit encoding in the body, but still
>> have it qp-encode headers that need it. Unfortunately, it seems I
>> can't do one without the other?
>Blargh.

>Unless somebody snuck an update to RFC2822 past me while I was out looking for
>my morning caffeine, non-7-bit-clean headers must *always* be encoded (either
>QP or =B or something), since there's no other way to pass the charset
>information for a non-ascii glyph.

>I haven't looked at the code yet, but this really should be refactored
>into two parts - one for the body and one for headers. I'm perfectly
>content to have the default for bodyparts be "none", since my MTA is
>able to do conversion as needed. But "none" isn't acceptable for the
>headers unless the MTA and exmh have a pre-existing agreement about
>what charset is in use...

>(And I'm going to LART the first person who points out that "none"
>really isn't an acceptable default for bodyparts because of the
>distinction between "binary" and "8bitmime" ;)

I just discovered another bug concerning this: the Fcc: header gets
qp-encoded along with the rest of the headers...

I just had a folder named "=?iso-8859-1?Q?Sl=E4kt?=" due to this.

//Christer

-- 
| Sys admin @ MEDIC              WWW: http://www.dtek.chalmers.se/            |
| Email: [email protected]   Phone: (0)31 772 5431, (0)707 53 57 57	      |
"I fought the loa and the loa won, I fought the loa and the loa won..."
        -- Dave Aronsson, a.s.r.