Quoted printable in multipart messages

Quoted printable in multipart messages

[Joakim Hove]

Hello,

I am struggling to get gnus to display Norwegian characters correctly
in quoted printable messages. Consider a raw message looking like
this:


   This is a multi-part message in MIME format.

   ------_=_NextPart_001_01C3A761.7570DF6C
   Content-Type: text/plain;
   	charset="iso-8859-1"
   Content-Transfer-Encoding: quoted-printable


   [...] Dr=F8bak.


   ....
   ....


When viewing this article it is displayed as:


Dro/bak


I.e. the norwegian character ø is decoded -> o/, better than nothing, 
but not really correct.

If I use the function quoted-printable-decode-region on the
particular string manually, it is rendered correctly. I guess this is
some setting which must be tuned - but ...?? If I get "normal" mail,
i.e. typically from another Unix user the Norwegian characters are
rendered correctly.


Any help greatly appreciated.

Joakim Hove


-- 
  /--------------------------------------------------------------------\
 / Joakim Hove  / hove@bccs.no  /  (55 5) 84076       |                 \
 | Unifob AS, Avdeling for Beregningsvitenskap (BCCS) | Stabburveien 18 |
 | CMU                                                | 5231 Paradis    |
 \ Thormøhlensgt.55, 5020 Bergen.                     | 55 91 28 18     /
  \--------------------------------------------------------------------/

Re: Quoted printable in multipart messages

[Kai Grossjohann]

Joakim Hove <hove@bccs.no> writes:

> I am struggling to get gnus to display Norwegian characters correctly
> in quoted printable messages. Consider a raw message looking like
> this:
>
>
>    This is a multi-part message in MIME format.
>
>    ------_=_NextPart_001_01C3A761.7570DF6C
>    Content-Type: text/plain;
>    	charset="iso-8859-1"
>    Content-Transfer-Encoding: quoted-printable
>
>
>    [...] Dr=F8bak.
>
>
>    ....
>    ....
>
>
> When viewing this article it is displayed as:
>
>
> Dro/bak
>
>
> I.e. the norwegian character ø is decoded -> o/, better than nothing, 
> but not really correct.

Well, you are using iso-8859-15 in your message, but the qp message
you quote is in iso-8859-1.  By default, Emacs assumes that ø in
iso-8859-1 and ø in iso-8859-15 are two different characters.

Maybe your terminal is configured to show iso-8859-15 characters (C-x
RET t)?  Not sure what's the implication for Emacs running under X11.
Hm.

I think starting in Emacs 21.3, iso-8859 unification is available and
(partly) turned on by default.  This might help.

Another thing is latin1-display.  This is a library that comes with
Emacs which tries to render non-X characters using approximations from
the X character set.  It seems that "o/" is an ascii approximation of
"ø".  So maybe you have that turned on?

Kai

Re: Quoted printable in multipart messages

[Joakim Hove]

Hello Kai,

thanks for answering.

> Well, you are using iso-8859-15 in your message, but the qp message
> you quote is in iso-8859-1.  

That is right - changing from 8859-1 -> 8859-15 was some desperate act
on my behalf, it probabbly only made things worse.


> I think starting in Emacs 21.3, iso-8859 unification is available and
> (partly) turned on by default.  This might help.

I am downloading Emacs 21.3 right now, with my fingers crossed!

> Another thing is latin1-display.  This is a library that comes with
> Emacs which tries to render non-X characters using approximations from
> the X character set.  It seems that "o/" is an ascii approximation of
> "ø".  

Yes - "o/" is a reasonable approximation to "ø", but as far as I can
see I am not using latin1-display:

   (message "Provided: %s" (featurep 'latin1-disp))

Gives nil. However, nevertheless I think you might have a point,
because the original text is rendered correctly with "ø" when I set
point and mark and call M-x quoted-printable-decode-region, indicating
that this might be something gnus is doing wrong?

Any sensible pointers to where one can learn more about things would
be nice.

Thanks - Joakim



-- 
  /--------------------------------------------------------------------\
 / Joakim Hove  / hove@bccs.no  /  (55 5) 84076       |                 \
 | Unifob AS, Avdeling for Beregningsvitenskap (BCCS) | Stabburveien 18 |
 | CMU                                                | 5231 Paradis    |
 \ Thormøhlensgt.55, 5020 Bergen.                     | 55 91 28 18     /
  \--------------------------------------------------------------------/

Re: Quoted printable in multipart messages

[Kai Grossjohann]

Joakim Hove <hove@bccs.no> writes:

> Yes - "o/" is a reasonable approximation to "ø", but as far as I can
> see I am not using latin1-display:
>
>    (message "Provided: %s" (featurep 'latin1-disp))
>
> Gives nil. However, nevertheless I think you might have a point,
> because the original text is rendered correctly with "ø" when I set
> point and mark and call M-x quoted-printable-decode-region, indicating
> that this might be something gnus is doing wrong?

I also vaguely remember that some functionality like this was included
in recent Oorts.  Does anyone know more?

> Any sensible pointers to where one can learn more about things would
> be nice.

Ah, moving the cursor over the "o/" would also be interesting -- is it
one char or two as far as C-f is concerned?

And then, C-u C-x = is interesting.

I'm guessing that Gnus is trying to be helpful, if it isn't
latin1-disp.

Kai

Re: Quoted printable in multipart messages

[Joakim Hove]

Hello Kai,

thanks for answering.

> Well, you are using iso-8859-15 in your message, but the qp message
> you quote is in iso-8859-1.  

That is right - changing from 8859-1 -> 8859-15 was some desperate act
on my behalf, it probabbly only made things worse.


> I think starting in Emacs 21.3, iso-8859 unification is available and
> (partly) turned on by default.  This might help.

I am downloading Emacs 21.3 right now, with my fingers crossed!

> Another thing is latin1-display.  This is a library that comes with
> Emacs which tries to render non-X characters using approximations from
> the X character set.  It seems that "o/" is an ascii approximation of
> "ø".  

Yes - "o/" is a reasonable approximation to "ø", but as far as I can
see I am not using latin1-display:

   (message "Provided: %s" (featurep 'latin1-disp))

Gives nil. However, nevertheless I think you might have a point,
because the original text is rendered correctly with "ø" when I set
point and mark and call M-x quoted-printable-decode-region, indicating
that this might be something gnus is doing wrong?

Any sensible pointers to where one can learn more about things would
be nice.

Thanks - Joakim



-- 
  /--------------------------------------------------------------------\
 / Joakim Hove  / hove@bccs.no  /  (55 5) 84076       |                 \
 | Unifob AS, Avdeling for Beregningsvitenskap (BCCS) | Stabburveien 18 |
 | CMU                                                | 5231 Paradis    |
 \ Thormøhlensgt.55, 5020 Bergen.                     | 55 91 28 18     /
  \--------------------------------------------------------------------/