10:40 a.m.
introspector is a package I've had in a sandbox for a while. It's
basically 21.5-ready, but I wanted to put it in xemacs-base and
xemacs-devel so it could be used with 21.4, too. introspector
consists of
- frame-utils.el, providing two functions to create frames with
minimal decorations, such as would be useful with tooltips and the
like. Unfortunately, under 21.4 I've found a focus problem (the tip
frame should not get keyboard focus, but it does) and a decoration
problem (I can't get the title bar to disappear). These probably
aren't show-stopperse, but they're horribly ugly.
- introspector.el, providing a popup version of `what-cursor-position'
which also provides more extensive information such as extents and
faces touching point. This is based on a general framework for
inspecting buffer objects.
I'm not sure how useful the general framework is, but the frame
utilities and the popup `buffer-position-introspector' are quite
useful as they are.
A git repository may be available later, but for now the package is
available as an XEmacs package at
http://turnbull.sk.tsukuba.ac.jp/Tools/XEmacs/introspector-0.80-pkg.tar.gz
Install it as usual. Sorry for the inconvenience, but you'll have to
explicitly load the introspector library. This does two invasive
things:
(1) it binds C-h = to `buffer-position-introspector', and
(2) adds a pre-command-hook to pop down the introspector popup on any
command.
Possibly this should be set up as a minor mode (comments welcome), and
then it would be no problem to have autoloads for the user-visible
entry points.
Testing and suggestions would be greatly appreciated. I think it
would be appropriate to put frame-utils in xemacs-base, if we can work
out the issues with 21.4. I'm of two minds about introspector, it
could go into xemacs-devel, but maybe it should go into xemacs-base
so that all users can use it to gather information for bug reports.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
2:49 a.m.
* Stephen J Turnbull <stephen(a)xemacs.org> writes:
- frame-utils.el, providing two functions to create frames with
minimal decorations, such as would be useful with tooltips and the
like. Unfortunately, under 21.4 I've found a focus problem (the tip
frame should not get keyboard focus, but it does) and a decoration
problem (I can't get the title bar to disappear).
Neither of these are a problem under SXEmacs. Lookie...
<http://www.sxemacs.org/~steve/tmp/funky.png>
- introspector.el
Haven't looked at this yet, but definitely will.
Install it as usual. Sorry for the inconvenience, but you'll
have to
explicitly load the introspector library. This does two invasive
things:
(1) it binds C-h = to `buffer-position-introspector', and
How about `C-h +'? It's currently undefined, and the `+' kinda suggests
"more info". Though I'm not sure that a key-binding is appropriate. I
understand that it is just a frame and you can pop up a new frame via
the keyboard, but this one gives the illusion of being a tooltip thing
or a dialog. Maybe a mouse binding would be better?
Possibly this should be set up as a minor mode (comments welcome),
and
then it would be no problem to have autoloads for the user-visible
entry points.
Good idea, I think.
Testing and suggestions would be greatly appreciated. I think it
would be appropriate to put frame-utils in xemacs-base, if we can work
out the issues with 21.4.
A couple of years back, you patched balloon-help so its frames would be
undecorated. I know that patch went into SXEmacs, but I'm not sure if
it made it into XEmacs 21.4. I bet you another beer that that patch
will fix frame-utils for 21.4. :-)
I'm of two minds about introspector, it could go into
xemacs-devel,
but maybe it should go into xemacs-base so that all users can use it
to gather information for bug reports.
I'd put them both in the same package. Doesn't really matter which,
-base or -devel seem equally appropriate to me.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
3:23 a.m.
2008/3/9 Steve Youngs <steve(a)sxemacs.org>:
* Stephen J Turnbull <stephen(a)xemacs.org> writes:
> Testing and suggestions would be greatly appreciated. I think it
> would be appropriate to put frame-utils in xemacs-base, if we can work
> out the issues with 21.4.
A couple of years back, you patched balloon-help so its frames would be
undecorated. I know that patch went into SXEmacs, but I'm not sure if
it made it into XEmacs 21.4. I bet you another beer that that patch
will fix frame-utils for 21.4. :-)
I'll check tomorrow, but I think that patch went in to 21.4.
- Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5:50 p.m.
Vin Shelton writes:
2008/3/9 Steve Youngs <steve(a)sxemacs.org>:
> * Stephen J Turnbull <stephen(a)xemacs.org> writes:
> > Testing and suggestions would be greatly appreciated. I think it
> > would be appropriate to put frame-utils in xemacs-base, if we can work
> > out the issues with 21.4.
>
> A couple of years back, you patched balloon-help so its frames would be
> undecorated. I know that patch went into SXEmacs, but I'm not sure if
> it made it into XEmacs 21.4. I bet you another beer that that patch
> will fix frame-utils for 21.4. :-)
I'll check tomorrow, but I think that patch went in to 21.4.
No, it's not there. It was a little difficult to find, so here's the
log
2005-04-27 Stephen J. Turnbull <stephen(a)xemacs.org>
Allow suppression of WM decorations.
* frame-x.c (Qoverride_redirect): New Lisp symbol.
(syms_of_frame_x): DEFSYMBOL it.
(x_create_widgets): Maybe set override_redirect property on frame.
(x_init_frame_1): Parse override-redirect from arguments.
(default-x-frame-plist): Document usage.
* frame.c (vars_of_frame): Improve docstring.
and diff:
Index: src/frame.c
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/src/frame.c,v
retrieving revision 1.70
retrieving revision 1.71
diff -u -r1.70 -r1.71
--- src/frame.c 2005/04/08 23:11:26 1.70
+++ src/frame.c 2005/04/27 09:01:48 1.71
＠＠ -3857,23 +3857,23 ＠＠
\(setq default-frame-plist '(width 80 height 55))
+Predefined properties are described in `set-frame-properties'.
+
The properties may be in alist format for backward compatibility
but you should not rely on this behavior.
These override values given in window system configuration data,
- including X Windows' defaults database.
+including X Windows' defaults database.
-Since the first X frame is created before loading your .emacs file,
-you must use the X resource database for that.
+Values for the first Emacs frame are taken from `initial-frame-plist'.
+Since the first X frame is created before loading your .emacs file, you
+may wish use the X resource database to avoid flashing.
-For values specific to the first Emacs frame, see `initial-frame-plist'.
For values specific to the separate minibuffer frame, see
- `minibuffer-frame-plist'.
-
-See also the variables `default-x-frame-plist' and
-`default-tty-frame-plist', which are like `default-frame-plist'
-except that they apply only to X or tty frames, respectively
-\(whereas `default-frame-plist' applies to all types of frames).
+`minibuffer-frame-plist'. See also the variables `default-x-frame-plist'
+and `default-tty-frame-plist', which are like `default-frame-plist'
+except that they apply only to X or tty frames, respectively \(whereas
+`default-frame-plist' applies to all types of frames).
*/ );
Vdefault_frame_plist = Qnil;
Index: src/frame-x.c
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/src/frame-x.c,v
retrieving revision 1.69
retrieving revision 1.70
diff -u -r1.69 -r1.70
--- src/frame-x.c 2005/01/24 23:33:56 1.69
+++ src/frame-x.c 2005/04/27 09:01:48 1.70
＠＠ -67,6 +67,7 ＠＠
/* Default properties to use when creating frames. */
Lisp_Object Vdefault_x_frame_plist;
+Lisp_Object Qoverride_redirect;
Lisp_Object Qx_resource_name;
static const struct memory_description x_frame_data_description_1 [] = {
＠＠ -1464,8 +1465,8 ＠＠
Widget app_shell = XtParent (wmshell);
Widget ew = FRAME_X_TEXT_WIDGET (f);
-/* set the position of the frame's root window now. When the
- frame was created, the position was initialized to (0,0). */
+ /* set the position of the frame's root window now. When the
+ frame was created, the position was initialized to (0,0). */
{
struct window *win = XWINDOW (f->root_window);
＠＠ -1770,13 +1771,15 ＠＠
/* Creates the widgets for a frame.
lisp_window_id is a Lisp description of an X window or Xt
widget to parse.
+ parent is a frame to use as the parent.
+ overridep if non-nil says to set the override-redirect setting.
This function does not create or map the windows. (That is
done by x_popup_frame().)
*/
static void
x_create_widgets (struct frame *f, Lisp_Object lisp_window_id,
- Lisp_Object parent)
+ Lisp_Object parent, Lisp_Object overridep)
{
struct device *d = XDEVICE (f->device);
Visual *visual = DEVICE_X_VISUAL (d);
＠＠ -1874,6 +1877,12 ＠＠
XtSetArg (al[ac], XtNcolormap, cmap); ac++;
}
+ if (!NILP (overridep))
+ {
+ XtSetArg (al[ac], XtNoverrideRedirect, True); ac++;
+ }
+
+ /* #### maybe we should check for FRAMEP instead? */
if (!NILP (parent))
{
parentwid = FRAME_X_SHELL_WIDGET (XFRAME (parent));
＠＠ -2085,6 +2094,7 ＠＠
Lisp_Object device = FRAME_DEVICE (f);
Lisp_Object lisp_window_id = Fplist_get (props, Qwindow_id, Qnil);
Lisp_Object popup = Fplist_get (props, Qpopup, Qnil);
+ Lisp_Object overridep = Fplist_get (props, Qoverride_redirect, Qnil);
if (!NILP (popup))
{
＠＠ -2109,7 +2119,7 ＠＠
f->visible = 1;
allocate_x_frame_struct (f);
- x_create_widgets (f, lisp_window_id, popup);
+ x_create_widgets (f, lisp_window_id, popup, overridep);
}
static void
＠＠ -2710,6 +2720,7 ＠＠
void
syms_of_frame_x (void)
{
+ DEFSYMBOL (Qoverride_redirect);
DEFSYMBOL (Qx_resource_name);
DEFSUBR (Fx_window_id);
＠＠ -2800,6 +2811,10 ＠＠
doing different things (e.g. not asking
for positioning, and not iconifying
separate from its parent).
+ override-redirect If non-nil, the frame will not be subject to
+ window-manager control. In particular, it
+ will lack decorations, for more attractive
+ appearance of balloon help, aka tooltips.
inter-line-space Not currently implemented.
toolbar-shadow-thickness Thickness of toolbar shadows.
background-toolbar-color Color of toolbar background.
==== end of diff ====
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:29 a.m.
Steve Youngs writes:
Neither of these are a problem under SXEmacs. Lookie...
<http://www.sxemacs.org/~steve/tmp/funky.png>
Cool! I didn't try too hard under 21.4, and I'm running under aquawm,
so I kinda expect strangeness here.
But, um, I notice that your xeyes is watching the clock like a hungry
hawk after a mouse. Do you Aussies work halfdays on Mondays? ;-)
> (1) it binds C-h = to `buffer-position-introspector', and
How about `C-h +'?
C-h = is by analogy to C-x =; in fact I'm kinda thinking in terms of
making it a configurable option for C-x =.
Maybe a mouse binding would be better?
I thought about that, but I just didn't implement it yet.
A couple of years back, you patched balloon-help so its frames would
be
undecorated. I know that patch went into SXEmacs, but I'm not sure if
it made it into XEmacs 21.4. I bet you another beer that that patch
will fix frame-utils for 21.4. :-)
Actually, frame-utils is that patch on steroids. I should check
whether it's still needed. I think that probably frame-utils or
something like it should be used to implement ballon-help.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
9:10 a.m.
Steve Youngs writes:
it made it into XEmacs 21.4. I bet you another beer that that patch
will fix frame-utils for 21.4. :-)
By the way, I was in a hurry when I answerd Vin, but you're right, I
owe you another beer.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
9:48 a.m.
Stephen J. Turnbull writes:
- frame-utils.el, providing two functions to create frames with
minimal decorations, such as would be useful with tooltips and the
like. Unfortunately, under 21.4 I've found a focus problem (the tip
frame should not get keyboard focus, but it does) and a decoration
problem (I can't get the title bar to disappear).
The title bar is there under Windows (cygwin) and 21.5 too.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
11:28 p.m.
Mats Lidell writes:
Stephen J. Turnbull writes:
> - frame-utils.el, providing two functions to create frames with
> minimal decorations, such as would be useful with tooltips and the
> like. Unfortunately, under 21.4 I've found a focus problem (the tip
> frame should not get keyboard focus, but it does) and a decoration
> problem (I can't get the title bar to disappear).
The title bar is there under Windows (cygwin)
Somebody who knows Windows will have to deal with that.
and 21.5 too.
Not on my Debian system (fvwm), my Gentoo system (fluxbox), or either
of my Macs (fluxbox and aquawm, respectively). I don't see how it
could happen, as override_redirect is enforced by Xlib AFAIK.
What window manager do you use?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
3:24 a.m.
Stephen J. Turnbull wrote:
> > - frame-utils.el, providing two functions to create frames
with
> > minimal decorations, such as would be useful with tooltips and the
> > like. Unfortunately, under 21.4 I've found a focus problem (the tip
> > frame should not get keyboard focus, but it does) and a decoration
> > problem (I can't get the title bar to disappear).
>
> The title bar is there under Windows (cygwin)
Somebody who knows Windows will have to deal with that.
> and 21.5 too.
Not on my Debian system (fvwm), my Gentoo system (fluxbox), or either
of my Macs (fluxbox and aquawm, respectively). I don't see how it
could happen, as override_redirect is enforced by Xlib AFAIK.
override_redirect is enforced by the X server. If a window has that
flag, create/move/resize requests are handled directly by the X server
regardless of any redirects.
It doesn't strictly prevent the WM (or anything else) from reparenting
the window, but the WM won't get notified of its creation (and even if
it did, wouldn't attempt to reparent it; you wouldn't want tooltips
and popup menus getting decorated).
The first thing I would check would be whether the override-redirect
Lisp property is actually resulting in the window having the
override_redirect flag set.
Assuming that it isn't, the next step would be to check the value of
the XtNoverrideRedirect resource on the Shell widget with editres,
i.e. whether the problem is in the conversion from Lisp to Xt, or from
Xt to X.
[I don't have 21.5 installed here, so I can't immediately check this.]
--
Glynn Clements <glynn(a)gclements.plus.com>
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
10:41 a.m.
Glynn Clements writes:
Assuming that it isn't, the next step would be to check the value
of
the XtNoverrideRedirect resource on the Shell widget with editres,
i.e. whether the problem is in the conversion from Lisp to Xt, or from
Xt to X.
Mats? I can't reproduce in 21.5.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
9:06 p.m.
>>>> Stephen wrote:
Stephen> Mats? I can't reproduce in 21.5.
Sorry if I was unclear but I said cygwin and 21.5 meaning running 21.5
under cygwin. It works fine under Gentoo with 21.5.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
11:35 p.m.
Mats Lidell writes:
>>>>> Stephen wrote:
Stephen> Mats? I can't reproduce in 21.5.
Sorry if I was unclear but I said cygwin and 21.5 meaning running 21.5
under cygwin. It works fine under Gentoo with 21.5.
The miscommunication is at least half on my part. I can say that the
usual idiom I would use express this is "21.5 *on* Cygwin". That
would be clearer to me. Please don't make unusual effort to remember
it.
Ah, OK. Are you running an X version or the native Windows GUI? If
you are using an X server it should work.
Doing this under Windows involves different magic. Do dialogs have
title bars under Windows? If so, it's probably hopeless unless we get
somebody with Windows GUI programming experience. If they don't,
there may be a Lisp interface to suppress decorations.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
8:18 a.m.
Stephen J. Turnbull writes:
Ah, OK. Are you running an X version or the native Windows GUI? If
you are using an X server it should work.
Native.
Doing this under Windows involves different magic. Do dialogs have
title bars under Windows?
No they don't so there is hope!? ;-)
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5885
days inactive
5890
days old
12 comments
5 participants
participants (5)
-
Glynn Clements -
Stephen J. Turnbull -
Steve Youngs -
The XEmacs Package Smoketest -
Vin Shelton