4:52 a.m.
You've probably noticed that if you try to paste a Latin1 character
such as "£" from Mozilla to XEmacs, you get a question-mark instead.
This is a bug in Mozilla that I reported a while ago
(http://bugzilla.mozilla.org/show_bug.cgi?id=136004) but I got
tired of waiting around for them to fix it, so I kluged a workaround:
before asking for the 'STRING target, check for the 'text/unicode
target.
But, I don't know anything about Unicode, so what I did to convert
it to Latin1 was just to delete every other byte. I leave the right
implementation to those who know/care...
(defun x-insert-selection (&optional check-cutbuffer-p move-point-event)
"Insert the current selection into buffer at point."
(interactive "P")
(let* ((unicode-p nil)
(text (or (condition-case ()
(prog1 (x-get-selection nil 'text/unicode)
(setq unicode-p t))
(error ()))
(x-get-selection))))
(cond (move-point-event
(mouse-set-point move-point-event)
(push-mark (point)))
((interactive-p)
(push-mark (point))))
(let ((p (point)))
(insert text)
(if unicode-p
;; When mozilla gives us unicode strings, the Latin1 characters
;; are in the top 8 bits of a 16 bit character: the bottom 8 bits
;; are null. Delete them. This is not even close to being a
;; proper Unicode parser, but at least it lets me paste Latin1
;; characters from Mozilla 1.1b to XEmacs 21.1.
(save-excursion
(while (> (point) p)
(delete-char -1)
(forward-char -1))))
text)))
--
Jamie Zawinski
jwz(a)jwz.org http://www.jwz.org/
jwz(a)dnalounge.com http://www.dnalounge.com/
9:17 p.m.
On Tue, 2002-08-27 at 06:52, Jamie Zawinski wrote:
You've probably noticed that if you try to paste a Latin1
character
such as "£" from Mozilla to XEmacs, you get a question-mark instead.
[snip]
I see something similar when I paste eg. my name from Ximian Evolution
(1.03) to XEmacs. It looks like:
"Ville Skytt%/1iso8859-15ä"
in my XEmacs. Don't know if this is related to the Mozilla stuff or
where this bug is. In case it matters, I have (latin-unity-install) in
my init.el.
--
\/ille Skyttä
ville.skytta at xemacs.org
3:09 a.m.
>>>> "Ville" == Ville Skytt <Ville>
writes:
Ville> On Tue, 2002-08-27 at 06:52, Jamie Zawinski wrote:
> You've probably noticed that if you try to paste a Latin1
> character such as "£" from Mozilla to XEmacs, you get a
> question-mark instead.
Ville> [snip]
Ville> I see something similar when I paste eg. my name from
Ville> Ximian Evolution (1.03) to XEmacs. It looks like:
Ville> "Ville Skytt%/1iso8859-15ä"
<RANT>
Similar but totally different. This is not a "bug", this is an
XFree86 self-inflicted lobotomy.
The Compound Text Encoding Standard, XFree86 edition, specifically
forbids this usage. File a bug against Evolution, and cite Section 5
of CText where ISO-8859-15 is declared to be a standard encoding, and
Section 6 where use of the extended segment format is forbidden for
standard encodings.
It's not Evolution's fault; XF86 has conflicting standards in place.
But the only real hope of getting this reversed is to make life
uncomfortable for everyone who uses it. And it _should_ be reversed;
there's no question in my mind that those paragraphs were intended to
cover _all_ 8859 encodings as they were approved. Furthermore, the
existing standard _just works_ for ISO 8859-15 in Mule (and any other
app where ISO 2022 is properly implemented)!
We have (thank all the three-letter acronyms who preceded me) already
gotten the Xutf8* extension removed. The whole motivation for the
Xutf8* extension was to save about 4 bytes of ISO-2022 header, and
incidentally make us revise every usage of Xmb* in our code. Once
again, if you have properly implemented ISO 2022, using UTF-8 with the
Xmb* functions "just works."
BTW, GNU Emacs has already caved in to the broken paste method based
on misguided European insistence, so we'll have to do so too. But I
think I'll borrow a note from Jamie, and bitch about it the way we
bitch about broken keymaps. ;-)
</RANT>
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
My nostalgia for Icon makes me forget about any of the bad things. I don't
have much nostalgia for Perl, so its faults I remember. Scott Gilbert c.l.py
6:45 p.m.
On Wed, 2002-08-28 at 05:09, Stephen J. Turnbull wrote:
Ville> I see something similar when I paste eg. my name from
Ville> Ximian Evolution (1.03) to XEmacs. It looks like:
Ville> "Ville Skytt%/1iso8859-15ä"
<RANT>
Similar but totally different. This is not a "bug", this is an
XFree86 self-inflicted lobotomy.
The Compound Text Encoding Standard, XFree86 edition, specifically
forbids this usage. File a bug against Evolution, and cite Section 5
of CText where ISO-8859-15 is declared to be a standard encoding, and
Section 6 where use of the extended segment format is forbidden for
standard encodings.
[snip]
Ok, thanks for the info, I just filed an Evolution bug about this in
Ximian bugzilla.
--
\/ille Skyttä
ville.skytta at xemacs.org
7916
days inactive
7917
days old
3 comments
3 participants
participants (3)
-
Jamie Zawinski -
Stephen J. Turnbull -
Ville Skyttä