utf-7 decoding?

utf-7 decoding?

[Felix Lee]

I see occasional messages that are utf-7 encoded without
appropriate mime headers and stuff.  Is there an easy way of
unmutilating them?  "C-u W M c" doesn't have a utf-7 thingy.

(utf-7 encoding looks like +ACI-this+ACI-)
--

Re: utf-7 decoding?

[François Pinard]

Felix Lee <flee@teleport.com> writes:

> I see occasional messages that are utf-7 encoded without
> appropriate mime headers and stuff.  Is there an easy way of
> unmutilating them?  "C-u W M c" doesn't have a utf-7 thingy.

For what it is worth, the next release of my `recode' tool does UTF-7.
Maybe there is some Emacs LISP code floating around to do the same?
If not, maybe the UTF-7 routines in `recode' could be recycled into Emacs
LISP code in Gnus?  I do not think that every such of those things should
be made part of Emacs guts in C, yet one never knows with Emacs :-).

Byte manipulations in Emacs LISP is usually rather slow.  For Q-P and Base64
conversions, I use `recode' instead of the Emacs LISP code (which is quite
natural as I'm exercising the pretest) and it is surely faster for me.

-- 
François Pinard                            mailto:pinard@iro.umontreal.ca
Join the free Translation Project!    http://www.iro.umontreal.ca/~pinard

Re: utf-7 decoding?

[Lars Magne Ingebrigtsen]

François Pinard <pinard@iro.umontreal.ca> writes:

> Byte manipulations in Emacs LISP is usually rather slow.  For Q-P and Base64
> conversions, I use `recode' instead of the Emacs LISP code (which is quite
> natural as I'm exercising the pretest) and it is surely faster for me.

The base64 functions are in the C layer in Emacs 20.4, so that should
be fast.  (The code is from recode, even.  :-)  Kind of.)  QP is
usually used when there's not all that much to encode (I mean, when
most of the text is us-ascii), so putting that into the C layer is
probably not worth the bother, I think.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: utf-7 decoding?

[François Pinard]

Lars Magne Ingebrigtsen <larsi@gnus.org> écrit:

> François Pinard <pinard@iro.umontreal.ca> writes:

> > Byte manipulations in Emacs LISP is usually rather slow.  For Q-P and Base64
> > conversions, I use `recode' instead of the Emacs LISP code (which is quite
> > natural as I'm exercising the pretest) and it is surely faster for me.

> The base64 functions are in the C layer in Emacs 20.4, so that should
> be fast.  (The code is from recode, even.  :-)  Kind of.)  QP is
> usually used when there's not all that much to encode (I mean, when
> most of the text is us-ascii), so putting that into the C layer is
> probably not worth the bother, I think.

I have a similar feeling.  And to me, UTF-7 is a kind of QP meant for
Unicode, and maybe a similar reasonning could make Emacs-LISP tolerable.
Instead of speaking about it, I should attempt writing some code.
Things should become a little less stressing for me in a few weeks, and
it might even be fun writing this.  But of course, Lars, I would not like
to steel this pleasure from you, if you feel like having it sooner! :-)

-- 
François Pinard                            mailto:pinard@iro.umontreal.ca
Join the free Translation Project!    http://www.iro.umontreal.ca/~pinard

Re: utf-7 decoding?

[Lars Magne Ingebrigtsen]

François Pinard <pinard@IRO.UMontreal.CA> writes:

> But of course, Lars, I would not like to steel this pleasure from
> you, if you feel like having it sooner! :-)

I'm starting to see people posting stuff using utf-7, so it would be
nice to have some sort of support for that.  But how would one do that 
using the Mule framework?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: utf-7 decoding?

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> François Pinard <pinard@IRO.UMontreal.CA> writes:
> 
> > But of course, Lars, I would not like to steel this pleasure from
> > you, if you feel like having it sooner! :-)
> 
> I'm starting to see people posting stuff using utf-7, so it would be
> nice to have some sort of support for that.  But how would one do
> that using the Mule framework?

The latest XEmacs/Mule is supposed to have UTF-8 support (I'm not sure 
about UTF-7), but it needs some special initialization.  Bleah.

Re: utf-7 decoding?

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> The latest XEmacs/Mule is supposed to have UTF-8 support (I'm not sure 
> about UTF-7), but it needs some special initialization.  Bleah.

But how does it work?  Does `char-charset' on a utf-8 char say
`utf-8', or something?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: utf-7 decoding?

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Hrvoje Niksic <hniksic@srce.hr> writes:
> 
> > The latest XEmacs/Mule is supposed to have UTF-8 support (I'm not sure 
> > about UTF-7), but it needs some special initialization.  Bleah.
> 
> But how does it work?  Does `char-charset' on a utf-8 char say
> `utf-8', or something?

No, utf-8 is implemented as a coding-system, not a charset.  This
means that you can decode external data from utf-8 to internal format, 
and result in having ASCII, Latin-1, Latin-2, and other characters
inserted in the buffer.  Which is quite correct, I'd say.

Understanding the difference between a (Mule) charset and a (Mule)
coding-system is the first step to (Mule) enlightenment.

Re: utf-7 decoding?

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> No, utf-8 is implemented as a coding-system, not a charset.  This
> means that you can decode external data from utf-8 to internal format, 
> and result in having ASCII, Latin-1, Latin-2, and other characters
> inserted in the buffer.  Which is quite correct, I'd say.

Oh, right.  That's certainly ok by me.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen