2:31 a.m.
Earlier today I updated my 21.2 tree from CVS. When I next started I
was invited to upgrade my ~/.emacs file to ~/.xemacs/init.el and
custom.el. I don't know if that's pertinent to what follows or not.
Now, when attempting to send a forward in gnus, I find myself accosted
with the message
Signaling: (selection-conversion-error "selection owner couldn't convert"
CLIPBOARD _EMACS_INTERNAL)
get-selection-internal(CLIPBOARD _EMACS_INTERNAL)
xselect-kill-buffer-hook-1(CLIPBOARD)
xselect-kill-buffer-hook()
kill-buffer(#<buffer " *temp*<2>">)
byte-code("..." [buffer-name temp-buffer kill-buffer] 2)
mml-generate-mime-1((part (type . "text/plain") (contents . "On Thu, 13
Jul 2000 20:40:02 -0400, \"Pastor Robert White\" <xrredeem(a)erie.net>
said:\n\n > when I mentioned the date to bishop Spring, he was until that\n >
moment, being, among other things, computer challenged, unaware\n > of the
date.\n\nHe didn't know what the date of Epiphany is?\n\n-- \nMichael D. Harnois,
Redeemer Lutheran Church, Washburn, IA \nmdharnois(a)home.com
aa0bt(a)aa0bt.ampr.org \n \"Believe those who are seeking the truth. \n Doubt those
who find it.\" -- Andre Gide\n")))
mml-generate-mime()
message-encode-message-body()
message-send-mail(nil)
message-send-via-mail(nil)
message-send(nil)
message-send-and-exit(nil)
call-interactively(message-send-and-exit)
and thenceforth I am screwed. I can't even quit out of gnus.
3:51 a.m.
these two occurrences are unrelated. i'll fix the second soon.
Michael Harnois wrote:
Earlier today I updated my 21.2 tree from CVS. When I next started I
was invited to upgrade my ~/.emacs file to ~/.xemacs/init.el and
custom.el. I don't know if that's pertinent to what follows or not.
Now, when attempting to send a forward in gnus, I find myself accosted
with the message
Signaling: (selection-conversion-error "selection owner couldn't convert"
CLIPBOARD _EMACS_INTERNAL)
get-selection-internal(CLIPBOARD _EMACS_INTERNAL)
xselect-kill-buffer-hook-1(CLIPBOARD)
xselect-kill-buffer-hook()
kill-buffer(#<buffer " *temp*<2>">)
byte-code("..." [buffer-name temp-buffer kill-buffer] 2)
mml-generate-mime-1((part (type . "text/plain") (contents . "On Thu, 13
Jul 2000 20:40:02 -0400, \"Pastor Robert White\" <xrredeem(a)erie.net>
said:\n\n > when I mentioned the date to bishop Spring, he was until that\n >
moment, being, among other things, computer challenged, unaware\n > of the
date.\n\nHe didn't know what the date of Epiphany is?\n\n-- \nMichael D. Harnois,
Redeemer Lutheran Church, Washburn, IA \nmdharnois(a)home.com
aa0bt(a)aa0bt.ampr.org \n \"Believe those who are seeking the truth. \n Doubt those
who find it.\" -- Andre Gide\n")))
mml-generate-mime()
message-encode-message-body()
message-send-mail(nil)
message-send-via-mail(nil)
message-send(nil)
message-send-and-exit(nil)
call-interactively(message-send-and-exit)
and thenceforth I am screwed. I can't even quit out of gnus.
--
Ben
In order to save my hands, I am cutting back on my mail. I also write
as succinctly as possible -- please don't be offended. If you send me
mail, you _will_ get a response, but please be patient, especially for
XEmacs-related mail. If you need an immediate response and it is not
apparent in your message, please say so. Thanks for your understanding.
See also http://www.666.com/ben/chronic-pain/
4:57 a.m.
try this patch and see if it fixes things.
alastair, it looks like the selection code needs more work -- it's doing all
these bizarre weird things relating to the x code, and this needs a bit of a
rethink.
the tricky thing with the x code is that the selection data can be one of three
types: 8, 16, or 32 bit data; under windows, we only have 8-bit data. 16 and
32-bit data automatically gets byte-swapped as necessary to deal with
endian-ness issues, necessary because clients on different architectures might
request data from the same X server. the c code tries to support this by
allowing conses of 16-bit values [because xemacs lisp integers are less than 32
bits] and vectors of numbers etc. to pass long 32-bit or 16-bit data. imho this
is reasonable but there's also lots of other crap going on. what i think you
should do is come up with a clean and simple interface for converting 16 and 32
bit data in and out, document it, and fix up the C code so it does this and no
more. [and fix the lisp code to correspond] you should also make the x code do
the proper convert calls if possible.
btw i do my testing under X using cygwin and a windows-based x server such as
MiX. i can give you more info on this if you want; otherwise, you can probably
do virtually all the testing without actually needing a setup like this.
ben
Index: select.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/Attic/select.c,v
retrieving revision 1.1.2.3
diff -u -r1.1.2.3 select.c
--- select.c 2000/07/13 09:06:35 1.1.2.3
+++ select.c 2000/07/14 03:40:55
＠＠ -133,8 +133,10 ＠＠
/* We couldn't. Try for a generic type instead. */
value = assq_no_quit (target_type, Qnil);
-
- return convert_selection (selection_symbol, value, target_type, 0);
+
+ /* #### URK! This should convert out AND in, and we shouldn't
+ be doing all this bizarre frobbing in the C code. */
+ return convert_selection (selection_symbol, value, target_type, 1);
}
return Qnil;
Michael Harnois wrote:
Earlier today I updated my 21.2 tree from CVS. When I next started I
was invited to upgrade my ~/.emacs file to ~/.xemacs/init.el and
custom.el. I don't know if that's pertinent to what follows or not.
Now, when attempting to send a forward in gnus, I find myself accosted
with the message
Signaling: (selection-conversion-error "selection owner couldn't convert"
CLIPBOARD _EMACS_INTERNAL)
get-selection-internal(CLIPBOARD _EMACS_INTERNAL)
xselect-kill-buffer-hook-1(CLIPBOARD)
xselect-kill-buffer-hook()
kill-buffer(#<buffer " *temp*<2>">)
byte-code("..." [buffer-name temp-buffer kill-buffer] 2)
mml-generate-mime-1((part (type . "text/plain") (contents . "On Thu, 13
Jul 2000 20:40:02 -0400, \"Pastor Robert White\" <xrredeem(a)erie.net>
said:\n\n > when I mentioned the date to bishop Spring, he was until that\n >
moment, being, among other things, computer challenged, unaware\n > of the
date.\n\nHe didn't know what the date of Epiphany is?\n\n-- \nMichael D. Harnois,
Redeemer Lutheran Church, Washburn, IA \nmdharnois(a)home.com
aa0bt(a)aa0bt.ampr.org \n \"Believe those who are seeking the truth. \n Doubt those
who find it.\" -- Andre Gide\n")))
mml-generate-mime()
message-encode-message-body()
message-send-mail(nil)
message-send-via-mail(nil)
message-send(nil)
message-send-and-exit(nil)
call-interactively(message-send-and-exit)
and thenceforth I am screwed. I can't even quit out of gnus.
--
Ben
In order to save my hands, I am cutting back on my mail. I also write
as succinctly as possible -- please don't be offended. If you send me
mail, you _will_ get a response, but please be patient, especially for
XEmacs-related mail. If you need an immediate response and it is not
apparent in your message, please say so. Thanks for your understanding.
See also http://www.666.com/ben/chronic-pain/
11:58 a.m.
Ben Wing <ben(a)666.com> writes:
try this patch and see if it fixes things.
Unfortunately, it's not. I think it also causes problem
under Windows.
Try this patch. I'll apply this on Monday if no one
objects.
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.c (get_local_selection): Fix argument to assq_no_quit.
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.el (selection-converter-alist): Set up converter for X.
Index: select.el
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/lisp/select.el,v
retrieving revision 1.5.2.9
diff -u -r1.5.2.9 select.el
--- select.el 2000/07/13 09:06:26 1.5.2.9
+++ select.el 2000/07/14 10:52:29
＠＠ -659,8 +659,27 ＠＠
;; Types listed here can be selections foreign to XEmacs
(setq selection-converter-in-alist
- '((CF_TEXT . select-convert-from-cf-text)
- ))
+ '((TEXT . select-convert-to-text)
+ (STRING . select-convert-to-string)
+ (COMPOUND_TEXT . select-convert-to-compound-text)
+ (TARGETS . select-convert-to-targets)
+ (LENGTH . select-convert-to-length)
+ (DELETE . select-convert-to-delete)
+ (FILE_NAME . select-convert-to-filename)
+ (CHARACTER_POSITION . select-convert-to-charpos)
+ (SOURCE_LOC . select-convert-to-sourceloc)
+ (LINE_NUMBER . select-convert-to-lineno)
+ (COLUMN_NUMBER . select-convert-to-colno)
+ (OWNER_OS . select-convert-to-os)
+ (HOST_NAME . select-convert-to-host)
+ (USER . select-convert-to-user)
+ (CLASS . select-convert-to-class)
+ (NAME . select-convert-to-name)
+ (ATOM . select-convert-to-atom)
+ (INTEGER . select-convert-to-integer)
+ (_EMACS_INTERNAL . select-convert-to-identity)
+ ((CF_TEXT . select-convert-from-cf-text)
+ )))
;; Types listed here can be appended by own-selection
(setq selection-appender-alist
Index: select.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/Attic/select.c,v
retrieving revision 1.1.2.3
diff -u -r1.1.2.3 select.c
--- select.c 2000/07/13 09:06:35 1.1.2.3
+++ select.c 2000/07/14 10:52:12
＠＠ -132,9 +132,11 ＠＠
return value;
/* We couldn't. Try for a generic type instead. */
- value = assq_no_quit (target_type, Qnil);
-
- return convert_selection (selection_symbol, value, target_type, 0);
+ value = assq_no_quit (Qnil, value_list);
+ if (CONSP (value))
+ value = XCDR (value);
+
+ return convert_selection (selection_symbol, value, target_type, 0);
}
return Qnil;
--
Yoshiki Hayashi
2:19 p.m.
-----Original Message-----
From: Yoshiki Hayashi [mailto:yoshiki＠xemacs.org]
Sent: 14 July 2000 11:58
To: Ben Wing
Cc: Michael Harnois; xemacs-beta(a)xemacs.org; Alastair J. Houghton;
xemacs-patches(a)xemacs.org
Subject: Re: fucked in gnus
Ben Wing <ben(a)666.com> writes:
> try this patch and see if it fixes things.
Unfortunately, it's not. I think it also causes problem
under Windows.
Try this patch. I'll apply this on Monday if no one
objects.
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.c (get_local_selection): Fix argument to assq_no_quit.
Just noticed this. I was about to fix it myself ;-)
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.el (selection-converter-alist): Set up converter for X.
Are you *sure* this is the right thing to do? I think there
should be a set of select-convert-from-* functions in
selection-converter-in-alist... This will work as a stopgap,
but I'm not sure it's quite right.
I think I have an alternative patch that additionally fixes
a few other things; when I've finished it (it's almost
ready), I'll post it.
Regards,
Alastair.
____________________________________________________________
Alastair Houghton ajhoughton(a)lineone.net
2:48 p.m.
Yoshiki Hayashi wrote:
Ben Wing <ben(a)666.com> writes:
> try this patch and see if it fixes things.
Unfortunately, it's not. I think it also causes problem
under Windows.
Try this patch. I'll apply this on Monday if no one
objects.
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.c (get_local_selection): Fix argument to assq_no_quit.
2000-07-14 Yoshiki Hayashi <yoshiki(a)xemacs.org>
* select.el (selection-converter-alist): Set up converter for X.
--------------------------------------------------------------------------------
Index: select.el
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/lisp/select.el,v
retrieving revision 1.5.2.9
diff -u -r1.5.2.9 select.el
--- select.el 2000/07/13 09:06:26 1.5.2.9
+++ select.el 2000/07/14 10:52:29
＠＠ -659,8 +659,27 ＠＠
;; Types listed here can be selections foreign to XEmacs
(setq selection-converter-in-alist
- '((CF_TEXT . select-convert-from-cf-text)
- ))
+ '((TEXT . select-convert-to-text)
+ (STRING . select-convert-to-string)
+ (COMPOUND_TEXT . select-convert-to-compound-text)
+ (TARGETS . select-convert-to-targets)
+ (LENGTH . select-convert-to-length)
+ (DELETE . select-convert-to-delete)
+ (FILE_NAME . select-convert-to-filename)
+ (CHARACTER_POSITION . select-convert-to-charpos)
+ (SOURCE_LOC . select-convert-to-sourceloc)
+ (LINE_NUMBER . select-convert-to-lineno)
+ (COLUMN_NUMBER . select-convert-to-colno)
+ (OWNER_OS . select-convert-to-os)
+ (HOST_NAME . select-convert-to-host)
+ (USER . select-convert-to-user)
+ (CLASS . select-convert-to-class)
+ (NAME . select-convert-to-name)
+ (ATOM . select-convert-to-atom)
+ (INTEGER . select-convert-to-integer)
+ (_EMACS_INTERNAL . select-convert-to-identity)
+ ((CF_TEXT . select-convert-from-cf-text)
+ )))
please let's not introduce nasty code like this. the only converter necessary
to fix the gnus problem is for _EMACS_INTERNAL, which should be using a new
function called select-convert-from-identity.
;; Types listed here can be appended by own-selection
(setq selection-appender-alist
--------------------------------------------------------------------------------
Index: select.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/Attic/select.c,v
retrieving revision 1.1.2.3
diff -u -r1.1.2.3 select.c
--- select.c 2000/07/13 09:06:35 1.1.2.3
+++ select.c 2000/07/14 10:52:12
＠＠ -132,9 +132,11 ＠＠
return value;
/* We couldn't. Try for a generic type instead. */
- value = assq_no_quit (target_type, Qnil);
-
- return convert_selection (selection_symbol, value, target_type, 0);
+ value = assq_no_quit (Qnil, value_list);
+ if (CONSP (value))
+ value = XCDR (value);
+
+ return convert_selection (selection_symbol, value, target_type, 0);
}
return Qnil;
--------------------------------------------------------------------------------
--
Yoshiki Hayashi
--
Ben
In order to save my hands, I am cutting back on my mail. I also write
as succinctly as possible -- please don't be offended. If you send me
mail, you _will_ get a response, but please be patient, especially for
XEmacs-related mail. If you need an immediate response and it is not
apparent in your message, please say so. Thanks for your understanding.
See also http://www.666.com/ben/chronic-pain/
8900
days inactive
8900
days old
5 comments
4 participants
participants (4)
-
Alastair Houghton -
Ben Wing -
Michael Harnois -
Yoshiki Hayashi