Re: lynx-dev a ¿Question?

[Klaus Weide]

On Wed, 2 Feb 2000, Eduardo Chappa L. wrote:

> Hello,
> 
>   This is hopefully a small question. I was trying to read news with lynx
> (comp.mail.pine). It had the advantage, I thought, that I could reload to
> check for new messages. However I found that some subjects contain strange
> characters that are not displayed by lynx. Maybe you'll find some of them
> in the subject of this message, but if you don't see them, this is what
> they look like:
> 
> =?ISO-8859-1?Q?Bug_in_Ordered_Subject_sort=3F_=28was_
> 
> I could not find any way to transform them into something I could read, is
> it there any way to do it from within lynx?

There is code in Lynx to decode this MIME header encoding in some cases,
but it only applies for the Japanese charset ISO-2022-JP.

I had generalized that to work for other charsets, and for a while
the generalized version was being used in the development code (before
2.8, I think).  The code is still in HTMIME.c but commented out:

   #ifdef NOTDEFINED
   /*
   **      Generalized HTmmdecode for chartrans - K. Weide 1997-03-06
   */
   PUBLIC void HTmmdecode ARGS2(

Compare also HTNews.c where HTmmdecode gets called.

It seems there were some problems with it, I'm not sure which they
were, so the generelized versing got disabled.  Unfortunately I never
got around to investigate what was wrong with it.

Maybe someone wants to try that disabled code, by reversing the logic of
the #ifdef NOTUSED{_CHARTRANS}...  (but then it may work less well for
Japanese, and the unknown problem(s) mentioned above may appear).

All this only applies to encoded header fields, not to MIME encoding
of message bodies.  There is no code in Lynx to decode those AFAIK.

Apart from that, the line you gave looks kind of invalid - somewhat
as if parts of it had undergone encoding twice - and it isn't complete
(no closing '?=' sequence).  So even with the right code Lynx might
not be able to decode *that* example.


   Klaus