w3/Gnus/Emacs 20 charset decoding

w3/Gnus/Emacs 20 charset decoding

[Lars Magne Ingebrigtsen]

I've installed Emacs 20.4.93 and w3.  w3 seems to reliably make Emacs
dump core (nice to see that Gnus isn't the only package to do that :-), 
so I've backed down to Emacs 20.4 instead.

Anyway, the problem is with non-ascii characters in web-based groups.
It would be nice if w3 could avoid decoding the characters -- is there
a way to make it do that?  That is, some coding-system-for-read in w3
should be bound to `binary' by Gnus before calling the w3 functions,
but which ones?  Heeelp.

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

Re: w3/Gnus/Emacs 20 charset decoding

[William M. Perry]

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

> I've installed Emacs 20.4.93 and w3.  w3 seems to reliably make Emacs
> dump core (nice to see that Gnus isn't the only package to do that :-),
> so I've backed down to Emacs 20.4 instead.

Come to the bleeding edge lars!  20.5.92.9 is the way to go!  Come see the
violence inherent in the new redisplay. :)

> Anyway, the problem is with non-ascii characters in web-based groups.  It
> would be nice if w3 could avoid decoding the characters -- is there a way
> to make it do that?  That is, some coding-system-for-read in w3 should be
> bound to `binary' by Gnus before calling the w3 functions, but which
> ones?  Heeelp.

Since I'm in the midst of revamping Emacs/W3 and URL to use your mm and
mailcap code, you tell me.  Is there anything I can do to make this easier?

The URL package has always just used the 'binary' method.  Then the w3
parser and friends use the charsets.  I'm actually going to be looking at
how Gnus does internationalization, since the way w3 does it sucks.

-Bill P.

Re: w3/Gnus/Emacs 20 charset decoding

[Lars Magne Ingebrigtsen]

wmperry@aventail.com (William M. Perry) writes:

> Come to the bleeding edge lars!  20.5.92.9 is the way to go!  Come see the
> violence inherent in the new redisplay. :)

Sounds much too scary.  :-)

> Since I'm in the midst of revamping Emacs/W3 and URL to use your mm and
> mailcap code, you tell me.  Is there anything I can do to make this easier?
> 
> The URL package has always just used the 'binary' method.  Then the w3
> parser and friends use the charsets.

That's the easiest way of all for me.  :-)  I just had to make sure I
was in a unibyte buffer before calling `url-insert-file-contents', and
then everything worked as it should, as if by magic.

> I'm actually going to be looking at how Gnus does
> internationalization, since the way w3 does it sucks.

There are many similarities between a newsreader and a web browser,
these days -- at least, it's the same MIME stuff (charset
specifications, encoding and so on.)

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

Re: w3/Gnus/Emacs 20 charset decoding

[Alan Shutko]

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

> I've installed Emacs 20.4.93 and w3.  w3 seems to reliably make Emacs
> dump core (nice to see that Gnus isn't the only package to do that :-), 
> so I've backed down to Emacs 20.4 instead.

Shortly after the post of .93's availability, Kenichi Handa posted a
patch which fixes that.  (Yes, I spent this morning debugging things
before realizing I'd forgotten to apply it.)

> Anyway, the problem is with non-ascii characters in web-based groups.

Oh... maybe this is a different problem?  Hmmm....

-- 
Alan Shutko <ats@acm.org> - In a variety of flavors!
ASCII a stupid question, you get an EBCDIC answer.

Re: w3/Gnus/Emacs 20 charset decoding

[Lars Magne Ingebrigtsen]

Alan Shutko <ats@acm.org> writes:

> Oh... maybe this is a different problem?  Hmmm....

Oh, the crash was unrelated to the charset problems I had.

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