viewing raw article

viewing raw article

[Vladimir Volovich]

Hi!

when i view the raw article using C-u g, gnus used to show 8-bit
characters as \200, etc, but it had changed recently and gnus now
shows such characters as eight-bit-graphic charset which appears to
use the iso8859-1 encoded font.

was such a change intensional? it seems that it is better to show raw
octets like \200 than latin1 chars in raw articles, or can i tell
emacs to show eight-bit-graphic characters as \200 etc?

Best,
v.

Re: viewing raw article

[Kai Großjohann]

Vladimir Volovich <vvv@vsu.ru> writes:

> when i view the raw article using C-u g, gnus used to show 8-bit
> characters as \200, etc, but it had changed recently and gnus now
> shows such characters as eight-bit-graphic charset which appears to
> use the iso8859-1 encoded font.
>
> was such a change intensional? it seems that it is better to show raw
> octets like \200 than latin1 chars in raw articles, or can i tell
> emacs to show eight-bit-graphic characters as \200 etc?

The change was intentional; it enables some people to forget that
there are multiple languages in the world.  For them, Emacs just
sees random bytes and interprets them in the locale.  I think they
call this eight bit cleanness or something like that.  Shudder.

I have the following bit of Lisp:

;; Display characters from the charsets `eight-bit-control' and
;; `eight-bit-graphic' as octal numbers in `font-lock-warning-face'

(defun egoge-display-eight-bit-as-octal (&optional face)
  (let ((face-offset (if face
			 (* (face-id face)
			    (expt 2 19))
		       0))
	char)
    (dotimes (i 128)
      (setq char (+ i 128))
      (aset standard-display-table char
	    (vconcat (mapcar (lambda (c)
			       (+ face-offset c))
			     (format "\\%o" char)))))))

(eval-after-load "font-lock"
  '(egoge-display-eight-bit-as-octal 'font-lock-warning-face))

Snarfed from...  Hm.  Who is egoge?  Oliver Scholz, it seems.  If it
was somebody else, please accept my apologies.

-- 
This line is not blank.

Re: viewing raw article

[Dave Love]

kai.grossjohann@gmx.net (Kai Großjohann) writes:

> The change was intentional;

What change was that?

> it enables some people to forget that
> there are multiple languages in the world.  For them, Emacs just
> sees random bytes and interprets them in the locale.

I don't think that's what's happening.  I assume this is eight-bit in
a multibyte buffer, not unibyte mode.  (Using the display table like
that isn't actually necessary in unibyte mode anyhow.)

Re: viewing raw article

[Dave Love]

Vladimir Volovich <vvv@vsu.ru> writes:

> Hi!
>
> when i view the raw article using C-u g, gnus used to show 8-bit
> characters as \200, etc, but it had changed recently and gnus now
> shows such characters as eight-bit-graphic charset which appears to
> use the iso8859-1 encoded font.

Do you really mean \200?  That's not in 8859-1 or eight-bit-graphic,
though I suppose such a font might have a glyph there.  Unless you
change the display table, it's displayed in octal.

eight-bit-graphic (\240-\377) is displayed with glyphs from your
default font.  This is bogus, but we couldn't persuade rms to
change it a while back.  (That's nothing to do with Gnus.)

I recommend:

(when default-enable-multibyte-characters
  (dotimes (i 128)
    (aset standard-display-table (+ 128 i) nil)))

Re: viewing raw article

[Vladimir Volovich]

"DL" == Dave Love writes:

 >> when i view the raw article using C-u g, gnus used to show 8-bit
 >> characters as \200, etc, but it had changed recently and gnus now
 >> shows such characters as eight-bit-graphic charset which appears
 >> to use the iso8859-1 encoded font.

 DL> Do you really mean \200?  That's not in 8859-1 or
 DL> eight-bit-graphic, though I suppose such a font might have a
 DL> glyph there.  Unless you change the display table, it's displayed
 DL> in octal.

an article may include character \200 (128) in it's body: some
character sets, not based on iso8859-* (e.g. cp866, koi8-r, etc), do
contain characters in the range \200-\237, so \200 is also possible
in the messages, and gnus used to show them as \200.

though, characters in the range \240-\377 occur more often, and these
were previously displayed as e.g. \240, but now they are shown as
latin1 characters when previewing raw article, which is bogus.

 DL> eight-bit-graphic (\240-\377) is displayed with glyphs from your
 DL> default font.  This is bogus, but we couldn't persuade rms to
 DL> change it a while back.  (That's nothing to do with Gnus.)

yes, this is very confusing and bogus.

 DL> (when default-enable-multibyte-characters (dotimes (i 128) (aset
 DL> standard-display-table (+ 128 i) nil)))

hm - this didn't change the way 8-bit characters are shown in raw
articles...

Best,
v.

Re: viewing raw article

[Kai Großjohann]

Dave Love <d.love@dl.ac.uk> writes:

> kai.grossjohann@gmx.net (Kai Großjohann) writes:
>
>> The change was intentional;
>
> What change was that?

It used to be that people saw \234 for non-ascii characters where
Emacs didn't know which charset they came from.  Now Emacs will
(often) display a glyph that looks like a character from their native
charset.

The display of eight-bit-graphic was changed.

>> it enables some people to forget that
>> there are multiple languages in the world.  For them, Emacs just
>> sees random bytes and interprets them in the locale.
>
> I don't think that's what's happening.  I assume this is eight-bit in
> a multibyte buffer, not unibyte mode.  (Using the display table like
> that isn't actually necessary in unibyte mode anyhow.)

I wasn't talking about unibyte or multibyte buffers, I was talking
about something that's at a higher level of abstraction.
-- 
This line is not blank.

Re: viewing raw article

[Dave Love]

kai.grossjohann@gmx.net (Kai Großjohann) writes:

> It used to be that people saw \234 for non-ascii characters where
> Emacs didn't know which charset they came from.

??  Emacs can't not know what charset such characters are from.
That's eight-bit-control (in Emacs 21).

> Now Emacs will (often) display a glyph that looks like a character
> from their native charset.

The standard display table has always (as far as I know) displayed
such things the same way.  They will be rendered with whatever glyphs
are at that position in the default font.

> The display of eight-bit-graphic was changed.

When and where?  That's what we wanted, but by default

(aref standard-display-table 255)
  => [255]

in all versions at least >= 20.7 as far as I can tell.

> I wasn't talking about unibyte or multibyte buffers, I was talking
> about something that's at a higher level of abstraction.

I don't understand that.  What do you mean?

Re: viewing raw article

[Dave Love]

Vladimir Volovich <vvv@vsu.ru> writes:

> an article may include character \200 (128) in it's body: some
> character sets, not based on iso8859-* (e.g. cp866, koi8-r, etc), do
> contain characters in the range \200-\237, so \200 is also possible
> in the messages,

Of course, but how are they being displayed as 8859-1, since they're
not 8859-1 characters?

> and gnus used to show them as \200.

It still does for me, and I'd be surprised to see eight-bit-control
displayed another way.

> though, characters in the range \240-\377 occur more often, and these
> were previously displayed as e.g. \240, but now they are shown as
> latin1 characters when previewing raw article, which is bogus.

Assuming you're in a Russian locale, it's pretty odd if they are
displayed as 8859-1.

> hm - this didn't change the way 8-bit characters are shown in raw
> articles...

It seems to me that you are picking up some customization.  I don't
see anything in Gnus controlling such things (except in summary
lines).