Re: Hopefully final GPM patch...
25 years, 2 months
William M. Perry
William M. Perry
Jan Vroonhof <vroonhof(a)math.ethz.ch> writes:
> wmperry(a)aventail.com (William M. Perry) writes:
>
> > Yup, I've actually got xemacs/modules/gpm ready to go if anybody
> > wants.
>
> Is there way to use the same source file for this so that it can be
> decided at configure time? (not that it is that relevant at this time,
> because Linux supports modules anyway).
It does use the exact some source file.
> > But to use this the temacs binary MUST be linked in such a way to export
> > all symbols (-Wl,--export-dynamic under Linux, etc) since we modify 2
> > global variables (event_stream and tty_console_methods)
>
> Don't we do that already? I thought that was part of the module support.
It didn't seem to get put into the link flags here with a clean configure.
> You have more module experience than I do. Would it be more appropriate
> to introduce some wrappers to set these (and make the variables static?).
event_stream cannot be static, too many places muck around with them. I
don't think it is worth it. I'll try to figure out why the link flags
weren't set correctly, and submit a patch if necessary.
I'm attaching the modules/gpm/Makefile that I hacked up. Just copy
gpmevent.c over into the directly and you should be all set.
-Bill P.
- 1 participants
- 0 comments
Re: Hopefully final GPM patch...
25 years, 2 months
Jan Vroonhof
Jan Vroonhof
wmperry(a)aventail.com (William M. Perry) writes:
> Yup, I've actually got xemacs/modules/gpm ready to go if anybody
> wants.
Is there way to use the same source file for this so that it can be
decided at configure time?
(not that it is that relevant at this time, because Linux supports
modules anyway).
> But to use this the temacs binary MUST be linked in such a way to export
> all symbols (-Wl,--export-dynamic under Linux, etc) since we modify 2
> global variables (event_stream and tty_console_methods)
Don't we do that already? I thought that was part of the module
support.
You have more module experience than I do. Would it be more
appropriate to introduce some wrappers to set these (and make the
variables static?).
Jan
- 1 participants
- 0 comments
Re: Hopefully final GPM patch...
25 years, 2 months
William M. Perry
William M. Perry
Jan Vroonhof <vroonhof(a)math.ethz.ch> writes:
> wmperry(a)aventail.com (William M. Perry) writes:
>
> > This incorporates all of ben's recent suggestions.
>
> Thanks. I'll try it out. I'll try to a few of the reported failure modes
> of the old implementation to see whether they are gone with this cleaner
> version. Do you have an idea whether it should now be possible to turn
> the GPM support into a module? From a first glance it looks that way.
Yup, I've actually got xemacs/modules/gpm ready to go if anybody wants.
But to use this the temacs binary MUST be linked in such a way to export
all symbols (-Wl,--export-dynamic under Linux, etc) since we modify 2
global variables (event_stream and tty_console_methods)
> P.S. How can this be final patch if there are no ChangeLogs? :-)
I'll send those along in a separate message later today. Have to get my
son ready to go to a birthday party, and we haven't gotten the present
yet. :)
-bp
- 1 participants
- 0 comments
Re: Hopefully final GPM patch...
25 years, 2 months
Jan Vroonhof
Jan Vroonhof
wmperry(a)aventail.com (William M. Perry) writes:
> This incorporates all of ben's recent suggestions.
Thanks. I'll try it out. I'll try to a few of the reported failure
modes of the old implementation to see whether they are gone with this
cleaner version. Do you have an idea whether it should now be possible
to turn the GPM support into a module? From a first glance it looks
that way.
Jan
P.S. How can this be final patch if there are no ChangeLogs? :-)
- 1 participants
- 0 comments
Re: Give people with crashes more of a clue
25 years, 2 months
Robin S. Socha
Robin S. Socha
* Hrvoje Niksic <hniksic(a)iskon.hr> writes:
> "Robin S. Socha" <robin(a)socha.net> writes:
>> * Hrvoje Niksic <hniksic(a)iskon.hr> writes:
>> Would it make any sense at all to not call the file PROBLEMS but
>> PROBLEMS.<your_os>
> Many items in PROBLEMS are OS-independent.
[...]
>> I remember seeing intimidated users that only knew KDE. Yes, I know
>> they should be using the RPMs.
> Huh?
Like you said: "Many items in PROBLEMS are OS-independent." So the file
tells a novice user whose sole conception of, say, Linux is that of
"the better Windows" about Alpha Linux ("is that a new distib?"), pgcc
("huh?"), AIX and other things Joe Random Linux-Poweruser ("look, Ma, I
found a DeadRat") has never heard of. I was trying to say that a file
that says "if you're using Linux or a SPARC, *read* me" is more likely
to be read. I recall reading the problems file months after my first
install, mainly because this kind of file is not too common in the
Linux world.
> PROBLEMS is supposed to be primarily a helper file for people who
> build XEmacs. Quite possibly some of the user-level documentation has
> leaked there, but I don't think it's sufficient to warrant a place at
> the menubar.
Is there reliable statistic about the XEmacs userbase? I may be very
wrong, but I would imagine that a large (and rapidly growing) number of
user are their own admins (ie single user systems running Linux or
*BSD). And why do I need a "Splash" entry in the Help menu?
--
Robin S. Socha
Big. Black. Beautified. <http://socha.net/XEmacs/>
- 1 participants
- 0 comments
Re: GPM mouse tracking issues solved, kindof
25 years, 2 months
William M. Perry
William M. Perry
Ben Wing <ben(a)666.com> writes:
> Some comments:
>
> -- Your code to enable GPM is definitely wrong. It will fail with run-temacs, when you start on an X device and later
> connect to a TTY, etc. You need to hook into create-device-hook.
Like this?
(defun gpm-create-device-hook (device)
(if (and (not noninteractive) ; Don't want to do this in batch mode
(fboundp 'gpm-enable) ; Must have C-level GPM support
(eq system-type 'linux) ; Must be running linux
(eq (device-type device) 'tty) ; on a tty
(equal "linux" (console-tty-terminal-type ; an a linux terminal type
(device-console device))))
(gpm-minor-mode 5)))
(add-hook 'create-device-hook 'gpm-create-device-hook)
> -- Your next_event callback does not honor user_p, like it should. (Motion events are not user events.)
Fixed... I'll be sending a final patch to xemacs-beta & xemacs-patches
later tonight or tomorrow morning.
> -- Replace XXX, ### and FIXME with ####.
Done as well.
> >The attached patch seems to work, and was really easy to add. My only
> >question is should 'real' events get priority? Right now, gpm events get
> >flushed first - will this screw up anything?
>
> Probably the events need to get ordered properly, but this is a hassle,
> and probably not worth it, esp. on TTY's where you can't have focus
> problems. But put a #### comments about this.
Will do.
-bp
- 1 participants
- 0 comments
Re: GPM mouse tracking issues solved, kindof
25 years, 2 months
William M. Perry
William M. Perry
Ben Wing <ben(a)666.com> writes:
> I didn't receive it twice, but I don't care about receiving things twice
> in any case.
>
> But you shouldn't be doing what you're doing. The event-stream code has
> many places that assume that only specific types of events (i.e. not
> mouse motion events) are on the command-event queue. We could modify
> such code to allow for this possibility, but it seems as though you might
> run into the "excessive numbers of motion events" problem -- i.e. if
> XEmacs is very slow and behind in processing events, the queue could fill
> up with large numbers of GPM motion events. With X or MSWIN, this won't
> happen because XEmacs won't ask for an event until it's ready for one,
> and X and MSWIN won't generate mouse events until they're requested.
>
> So I think the cleanest way would be to do exactly what Jan proposes --
> don't try to stuff your events at all; instead, create your own queue for
> motion, button, and all other GPM events, and put wrapper functions
> around the next-event and event-pending-p callbacks.
>
> When you do this, however, make sure you grep through the code for all
> places that might change these callbacks -- I have a vague feeling that
> in some cases (e.g. some debug-enabled code?), the callbacks might get
> changed in the middle of a running XEmacs.
It doesn't look like anybody (but me now) changes event_stream once it is
created.
The attached patch seems to work, and was really easy to add. My only
question is should 'real' events get priority? Right now, gpm events get
flushed first - will this screw up anything?
The only thing that doesn't work right now is multiple GPM-aware devices
within the same running XEmacs. I tried to get this working, but multiple
calls to Gpm_Open don't seem to work. My basic approach was going to be:
1) gpm-enable would call Gpm_Open
2) If it succeeded, it would create the subprocess object
3) reset global gpm_fd to -1
4) reset global gpm_tried to 0, so that Gpm_Open will retry opening its
supposedly 'dead' socket.
5) reset gpm_fd just before the Gpm_GetEvent() call in Freceive_gpm_event
by grabbing the input fd from the process object, and set it back to -1
afterwards.
Some of the internal caching that Gpm_Open does screws this up, I'm not
sure why. I'm going to call this good enough for now. If I can get
ViaVoice running on my desktop or my laptop I might start playing around
with converting that to the connect_to_file_descriptor() method of things.
Anybody played with ViaVoice OutLoud yet? You could make a speech
synthesizer for EmacsSpeak that didnt' require a separate synthesizer. You
could hook right into redisplay...
-Bill P.
Index: lisp/gpm.el
*** gpm.el Tue May 5 15:32:27 1998
--- gpm.el Fri Nov 5 13:57:23 1999
***************
*** 0 ****
--- 1,70 ----
+ ;;; gpm.el --- Support the mouse when emacs run on a Linux console.
+
+ ;; Copyright (C) 1999 Free Software Foundation
+
+ ;; Author: William Perry <wmperry(a)gnu.org>
+ ;; Keywords: mouse, terminals
+
+ ;; This file is part of XEmacs.
+
+ ;; XEmacs is free software; you can redistribute it and/or modify
+ ;; it under the terms of the GNU General Public License as published by
+ ;; the Free Software Foundation; either version 2, or (at your option)
+ ;; any later version.
+
+ ;; XEmacs is distributed in the hope that it will be useful,
+ ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+ ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ ;; GNU General Public License for more details.
+
+ ;; You should have received a copy of the GNU General Public License
+ ;; along with GNU Emacs; see the file COPYING. If not, write to the
+ ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ ;; Boston, MA 02111-1307, USA.
+
+ (defvar gpm-minor-mode nil
+ "Non-nil when gpm-minor-mode is active. Call `gpm-minor-mode' to toggle.")
+
+ (defun gpm-minor-mode (&optional arg)
+ "Toggle GPM mouse mode.
+ With prefix arg, turn GPM mouse mode on if and only if arg is positive."
+ (interactive "P")
+ (cond
+ ((null arg) ; Toggle
+ (if gpm-minor-mode
+ (progn
+ (gpm-enable nil)
+ (setq gpm-minor-mode nil))
+ (if (gpm-enable t)
+ (setq gpm-minor-mode t))))
+ ((> arg 0) ; Turn on
+ (if (gpm-enable t)
+ (setq gpm-minor-mode t)))
+ (t ; Turn off
+ (gpm-enable nil)
+ (setq gpm-minor-mode nil))))
+
+ (defun turn-on-gpm-mouse-tracking ()
+ ;; Enable mouse tracking on linux console
+ (if gpm-minor-mode
+ (gpm-enable t)))
+
+ (defun turn-off-gpm-mouse-tracking ()
+ ;; Disable mouse tracking on linux console
+ (if gpm-minor-mode
+ (gpm-enable nil)))
+
+ ;; Restore normal mouse behaviour outside Emacs
+
+ (add-hook 'suspend-hook 'turn-off-gpm-mouse-tracking)
+ (add-hook 'suspend-resume-hook 'turn-on-gpm-mouse-tracking)
+ (add-hook 'kill-emacs-hook 'turn-off-gpm-mouse-tracking)
+
+ (if (and (not noninteractive) ; Don't want to do this in batch mode
+ (fboundp 'gpm-enable) ; Must have C-level GPM support
+ (eq system-type 'linux) ; Must be running linux
+ (eq (device-type) 'tty) ; on a tty
+ (equal "linux" (getenv "TERM"))) ; and it must be a linux virtual console
+ (gpm-minor-mode 5))
+
+ (provide 'gpm)
Index: lisp/mouse.el
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/lisp/mouse.el,v
retrieving revision 1.19.2.6
diff -u -r1.19.2.6 mouse.el
--- mouse.el 1999/09/02 08:47:55 1.19.2.6
+++ mouse.el 1999/11/06 12:07:37
＠＠ -88,7 +88,9 ＠＠
the X cutbuffer. A mark is pushed, so that the inserted text lies
between point and mark."
(interactive)
- (if (not (console-on-window-system-p))
+ (if (and (not (console-on-window-system-p))
+ (and (featurep 'gpm)
+ (not gpm-minor-mode)))
(yank)
(push-mark)
(if (region-active-p)
Index: src/console-tty.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/console-tty.c,v
retrieving revision 1.18.2.5
diff -u -r1.18.2.5 console-tty.c
--- console-tty.c 1999/10/24 03:48:31 1.18.2.5
+++ console-tty.c 1999/11/06 12:07:39
＠＠ -38,9 +38,6 ＠＠
#ifdef FILE_CODING
#include "file-coding.h"
#endif
-#ifdef HAVE_GPM
-#include "gpmevent.h"
-#endif
DEFINE_CONSOLE_TYPE (tty);
DECLARE_IMAGE_INSTANTIATOR_FORMAT (nothing);
＠＠ -129,10 +126,6 ＠＠
#endif /* MULE */
tty_con->terminal_type = terminal_type;
tty_con->controlling_process = controlling_process;
-
-#ifdef HAVE_GPM
- connect_to_gpm (con);
-#endif
if (NILP (CONSOLE_NAME (con)))
CONSOLE_NAME (con) = Ffile_name_nondirectory (tty);
Index: src/console-tty.h
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/console-tty.h,v
retrieving revision 1.8.2.2
diff -u -r1.8.2.2 console-tty.h
--- console-tty.h 1999/02/17 20:03:57 1.8.2.2
+++ console-tty.h 1999/11/06 12:07:39
＠＠ -41,9 +41,6 ＠＠
struct tty_console
{
int infd, outfd;
-#ifdef HAVE_GPM
- int mouse_fd;
-#endif
Lisp_Object instream, outstream;
Lisp_Object terminal_type;
Lisp_Object controlling_process;
＠＠ -203,9 +200,6 ＠＠
unsigned int is_stdio :1;
};
-#ifdef HAVE_GPM
-#define CONSOLE_TTY_MOUSE_FD(c) (CONSOLE_TTY_DATA (c)->mouse_fd)
-#endif
#define CONSOLE_TTY_DATA(c) CONSOLE_TYPE_DATA (c, tty)
#define CONSOLE_TTY_CURSOR_X(c) (CONSOLE_TTY_DATA (c)->cursor_x)
#define CONSOLE_TTY_CURSOR_Y(c) (CONSOLE_TTY_DATA (c)->cursor_y)
Index: src/device-tty.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/device-tty.c,v
retrieving revision 1.9
diff -u -r1.9 device-tty.c
--- device-tty.c 1998/05/27 20:42:00 1.9
+++ device-tty.c 1999/11/06 12:07:39
＠＠ -38,10 +38,6 ＠＠
#include "syssignal.h" /* for SIGWINCH */
-#ifdef HAVE_GPM
-#include <gpm.h>
-#endif
-
#include <errno.h>
Lisp_Object Qinit_pre_tty_win, Qinit_post_tty_win;
＠＠ -155,15 +151,6 ＠＠
CONSOLE_TTY_DATA (con)->width = width;
CONSOLE_TTY_DATA (con)->height = height;
-#ifdef HAVE_GPM
- /* We need to tell GPM how big our screen is now
- ** I am pretty sure the GPM library will get incredibly confused
- ** if you try to connect to more than one mouse-capable device,
- ** so I don't think it will cause any more damage in that case.
- */
- gpm_mx = width;
- gpm_my = height;
-#endif
for (tail = DEVICE_FRAME_LIST (d);
!NILP (tail);
tail = XCDR (tail))
Index: src/emacs.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/emacs.c,v
retrieving revision 1.82.2.32
diff -u -r1.82.2.32 emacs.c
--- emacs.c 1999/11/02 02:41:45 1.82.2.32
+++ emacs.c 1999/11/06 12:07:44
＠＠ -1088,6 +1088,10 ＠＠
syms_of_eldap ();
#endif
+#ifdef HAVE_GPM
+ syms_of_gpmevent ();
+#endif
+
/* Now create the subtypes for the types that have them.
We do this before the vars_*() because more symbols
may get initialized here. */
＠＠ -1481,6 +1485,10 ＠＠
#ifdef HAVE_LDAP
vars_of_eldap ();
+#endif
+
+#ifdef HAVE_GPM
+ vars_of_gpmevent ();
#endif
/* Now initialize any specifier variables. We do this later
Index: src/event-Xt.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/event-Xt.c,v
retrieving revision 1.41.2.14
diff -u -r1.41.2.14 event-Xt.c
--- event-Xt.c 1999/10/10 11:45:15 1.41.2.14
+++ event-Xt.c 1999/11/06 12:07:46
＠＠ -2034,9 +2034,6 ＠＠
{
Lisp_Object console;
int infd;
-#ifdef HAVE_GPM
- int mousefd;
-#endif
if (CONSOLE_X_P (con))
return; /* X consoles are automatically selected for when we
＠＠ -2044,22 +2041,6 ＠＠
infd = event_stream_unixoid_select_console (con);
XSETCONSOLE (console, con);
select_filedesc (infd, console);
-#ifdef HAVE_GPM
- /* On a stream device (ie: noninteractive), bad things can happen. */
- if (EQ (CONSOLE_TYPE (con), Qtty)) {
- mousefd = CONSOLE_TTY_MOUSE_FD (con);
- /* We check filedesc_to_what_closure[fd] here because if you run
- ** XEmacs from a TTY, it will fire up GPM, select the mouse fd, then
- ** if you run gnuattach to connect to another TTY, it will fire up
- ** GPM again, and try to reselect the mouse fd. GPM uses the same
- ** fd for every connection apparently, and select_filedesc will
- ** fail its assertion if we try to select it twice.
- */
- if ((mousefd >= 0) && !filedesc_to_what_closure[mousefd]) {
- select_filedesc (mousefd, console);
- }
- }
-#endif
}
static void
＠＠ -2067,9 +2048,6 ＠＠
{
Lisp_Object console;
int infd;
-#ifdef HAVE_GPM
- int mousefd;
-#endif
if (CONSOLE_X_P (con))
return; /* X consoles are automatically selected for when we
＠＠ -2077,15 +2055,6 ＠＠
infd = event_stream_unixoid_unselect_console (con);
XSETCONSOLE (console, con);
unselect_filedesc (infd);
-#ifdef HAVE_GPM
- /* On a stream device (ie: noninteractive), bad things can happen. */
- if (EQ (CONSOLE_TYPE (con), Qtty)) {
- mousefd = CONSOLE_TTY_MOUSE_FD (con);
- if (mousefd >= 0) {
- unselect_filedesc (mousefd);
- }
- }
-#endif
}
/* read an event from a tty, if one is available. Returns non-zero
Index: src/event-unixoid.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/event-unixoid.c,v
retrieving revision 1.9.2.1
diff -u -r1.9.2.1 event-unixoid.c
--- event-unixoid.c 1999/03/07 15:25:17 1.9.2.1
+++ event-unixoid.c 1999/11/06 12:07:53
＠＠ -41,10 +41,6 ＠＠
#include "sysproc.h" /* select stuff */
#include "systime.h"
-#ifdef HAVE_GPM
-#include "gpmevent.h"
-#endif
-
/* Mask of bits indicating the descriptors that we wait for input on.
These work as follows:
＠＠ -83,12 +79,6 ＠＠
Lisp_Object console;
XSETCONSOLE (console, con);
-
-#ifdef HAVE_GPM
- if (fd == CONSOLE_TTY_MOUSE_FD (con)) {
- return handle_gpm_read (event,con,fd);
- }
-#endif
nread = read (fd, &ch, 1);
if (nread <= 0)
Index: src/frame-tty.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/frame-tty.c,v
retrieving revision 1.8.2.1
diff -u -r1.8.2.1 frame-tty.c
--- frame-tty.c 1998/12/05 16:55:48 1.8.2.1
+++ frame-tty.c 1999/11/06 12:07:53
＠＠ -33,10 +33,6 ＠＠
#include "events.h"
-#ifdef HAVE_GPM
-#include <gpm.h>
-#endif
-
/* Default properties to use when creating frames. */
Lisp_Object Vdefault_tty_frame_plist;
＠＠ -86,33 +82,6 ＠＠
call1 (Qinit_post_tty_win, FRAME_CONSOLE (f));
}
-#ifdef HAVE_GPM
-static int
-tty_get_mouse_position (struct device *d, Lisp_Object *frame, int *x, int *y)
-{
- Gpm_Event ev;
- int num_buttons;
-
- num_buttons = Gpm_GetSnapshot(&ev);
- *x = ev.x;
- *y = ev.y;
- *frame = DEVICE_SELECTED_FRAME (d);
- return (1);
-}
-
-static void
-tty_set_mouse_position (struct window *w, int x, int y)
-{
- /* XXX
- I couldn't find any GPM functions that set the mouse position.
- Mr. Perry had left this function empty; that must be why.
- karlheg
- */
-}
-
-#endif
-
-
/* Change from withdrawn state to mapped state. */
static void
tty_make_frame_visible (struct frame *f)
＠＠ -223,10 +192,6 ＠＠
CONSOLE_HAS_METHOD (tty, init_frame_1);
CONSOLE_HAS_METHOD (tty, init_frame_3);
CONSOLE_HAS_METHOD (tty, after_init_frame);
-#ifdef HAVE_GPM
- CONSOLE_HAS_METHOD (tty, get_mouse_position);
- CONSOLE_HAS_METHOD (tty, set_mouse_position);
-#endif
CONSOLE_HAS_METHOD (tty, make_frame_visible);
CONSOLE_HAS_METHOD (tty, make_frame_invisible);
CONSOLE_HAS_METHOD (tty, frame_visible_p);
Index: src/gpmevent.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/gpmevent.c,v
retrieving revision 1.5
diff -u -r1.5 gpmevent.c
--- gpmevent.c 1998/03/31 20:11:47 1.5
+++ gpmevent.c 1999/11/06 12:07:53
＠＠ -1,5 +1,28 ＠＠
-/* William Perry 1997 */
+/* GPM functions
+ Copyright (C) 1997 William M. Perry <wmperry(a)gnu.org>
+ Copyright (C) 1999 Free Software Foundation, Inc.
+
+This file is part of XEmacs.
+
+XEmacs is free software; you can redistribute it and/or modify it
+under the terms of the GNU General Public License as published by the
+Free Software Foundation; either version 2, or (at your option) any
+later version.
+
+XEmacs is distributed in the hope that it will be useful, but WITHOUT
+ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with XEmacs; see the file COPYING. If not, write to
+the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA. */
+/* Synched up with: Not in FSF. */
+
+/* Authors: William Perry */
+
#include <config.h>
#include "lisp.h"
#include "console.h"
＠＠ -8,11 +31,18 ＠＠
#include "events.h"
#include "events-mod.h"
#include "sysdep.h"
+#include "commands.h"
+#include "lstream.h"
+#include "process.h"
#ifdef HAVE_GPM
#include "gpmevent.h"
#include <gpm.h>
+/* This is commented out in commands.h, I'm not sure why */
+extern Lisp_Object Vunread_command_events;
+extern Lisp_Object Vunread_command_event;
+
#if (!defined(__linux__)) /* possible under xterm */
#define KG_SHIFT 0
#define KG_CTRL 2
＠＠ -20,95 +50,405 ＠＠
#else
#include <linux/keyboard.h>
#endif
+
+static int (*orig_event_pending_p) (int);
+static void (*orig_next_event_cb) (struct Lisp_Event *);
+
+static Lisp_Object gpm_event_queue;
+static Lisp_Object gpm_event_queue_tail;
+
+static int
+get_process_infd (struct Lisp_Process *p)
+{
+ Lisp_Object instr, outstr;
+ get_process_streams (p, &instr, &outstr);
+ assert (!NILP (instr));
+ return filedesc_stream_fd (XLSTREAM (instr));
+}
+
+DEFUN ("receive-gpm-event", Freceive_gpm_event, 0, 2, 0, /*
+Run GPM_GetEvent().
+This function is the process handler for the GPM connection.
+*/
+ (process, string))
+{
+ Gpm_Event ev;
+ int modifiers = 0;
+ int type = -1;
+ int button = 1;
+ Lisp_Object fake_event;
+ Lisp_Process *proc;
+ struct Lisp_Event *event = NULL;
+ struct gcpro gcpro1;
+ static int num_events;
+
+ CHECK_PROCESS (process);
+
+ if (!Gpm_GetEvent(&ev))
+ {
+ warn_when_safe (Qnil, Qcritical, "Gpm_GetEvent failed");
+ return(Qzero);
+ }
+
+ GCPRO1(fake_event);
+
+ num_events++;
+
+ fake_event = Fmake_event (Qnil, Qnil);
+ event = XEVENT(fake_event);
+
+ event->timestamp = 0;
+ event->channel = Fselected_frame (Qnil); /* CONSOLE_SELECTED_FRAME (con); */
+
+ /* Whow, wouldn't named defines be NICE!?!?! */
+ modifiers = 0;
+
+ if (ev.modifiers & 1) modifiers |= MOD_SHIFT;
+ if (ev.modifiers & 2) modifiers |= MOD_META;
+ if (ev.modifiers & 4) modifiers |= MOD_CONTROL;
+ if (ev.modifiers & 8) modifiers |= MOD_META;
+
+ if (ev.buttons & GPM_B_LEFT)
+ {
+ button = 1;
+ }
+ else if (ev.buttons & GPM_B_MIDDLE)
+ {
+ button = 2;
+ }
+ else if (ev.buttons & GPM_B_RIGHT)
+ {
+ button = 3;
+ }
+
+ switch (GPM_BARE_EVENTS(ev.type)) {
+ case GPM_DOWN:
+ case GPM_UP:
+ event->event_type =
+ (ev.type & GPM_DOWN) ? button_press_event : button_release_event;
+ event->event.button.x = ev.x;
+ event->event.button.y = ev.y;
+ event->event.button.button = button;
+ event->event.button.modifiers = modifiers;
+ break;
+ case GPM_MOVE:
+ case GPM_DRAG:
+ event->event_type = pointer_motion_event;
+ event->event.motion.x = ev.x;
+ event->event.motion.y = ev.y;
+ event->event.motion.modifiers = modifiers;
+ default:
+ /* This will never happen */
+ break;
+ }
+
+ /* Handle the event */
+ enqueue_event (fake_event, &gpm_event_queue, &gpm_event_queue_tail);
+
+ GPM_DRAWPOINTER(&ev);
+ UNGCPRO;
+
+ return (Qzero);
+}
-int
-handle_gpm_read (struct Lisp_Event *event, struct console *con, int fd)
+static void turn_off_gpm (void)
{
- Gpm_Event ev;
- int modifiers = 0;
- int type = -1;
- int button = 1;
-
- if (!Gpm_GetEvent(&ev))
- return 0;
-
- event->timestamp = 0;
- event->channel = CONSOLE_SELECTED_FRAME (con);
-
- /* Whow, wouldn't named defines be NICE!?!?! */
- modifiers = 0;
-
- if (ev.modifiers & 1) modifiers |= MOD_SHIFT;
- if (ev.modifiers & 2) modifiers |= MOD_META;
- if (ev.modifiers & 4) modifiers |= MOD_CONTROL;
- if (ev.modifiers & 8) modifiers |= MOD_META;
-
- if (ev.type & GPM_DOWN)
- type = GPM_DOWN;
- else if (ev.type & GPM_UP)
- type = GPM_UP;
- else if (ev.type & GPM_MOVE) {
- type = GPM_MOVE;
- GPM_DRAWPOINTER(&ev);
- }
-
- if (ev.buttons & GPM_B_LEFT)
- button = 1;
- else if (ev.buttons & GPM_B_MIDDLE)
- button = 2;
- else if (ev.buttons & GPM_B_RIGHT)
- button = 3;
-
- switch (type) {
- case GPM_DOWN:
- case GPM_UP:
- event->event_type =
- type == GPM_DOWN ? button_press_event : button_release_event;
- event->event.button.x = ev.x;
- event->event.button.y = ev.y;
- event->event.button.button = button;
- event->event.button.modifiers = modifiers;
- break;
- case GPM_MOVE:
- event->event_type = pointer_motion_event;
- event->event.motion.x = ev.x;
- event->event.motion.y = ev.y;
- event->event.motion.modifiers = modifiers;
- default:
- return 0;
- }
- return 1;
-}
-
-void
-connect_to_gpm (struct console *con)
-{
- /* Only do this if we are running after dumping and really interactive */
- if (!noninteractive && initialized) {
- /* We really only want to do this on a TTY */
- CONSOLE_TTY_MOUSE_FD (con) = -1;
- if (EQ (CONSOLE_TYPE (con), Qtty)) {
- Gpm_Connect conn;
- int rval;
-
- conn.eventMask = GPM_DOWN|GPM_UP|GPM_MOVE;
- conn.defaultMask = GPM_MOVE;
- conn.minMod = 0;
- conn.maxMod = ((1<<KG_SHIFT)|(1<<KG_ALT)|(1<<KG_CTRL));
-
- rval = Gpm_Open (&conn, 0);
- switch (rval) {
- case -1: /* General failure */
- break;
- case -2: /* We are running under an XTerm */
+ Fdelete_process (build_string ("gpm"));
Gpm_Close();
- break;
- default:
- set_descriptor_non_blocking (gpm_fd);
- CONSOLE_TTY_MOUSE_FD (con) = gpm_fd;
- }
- }
- }
+
+ /* Remove the TTY mouse methods */
+ /* This is breaking abstraction a bit, but... */
+ tty_console_methods->get_mouse_position_method = NULL;
+ tty_console_methods->set_mouse_position_method = NULL;
+ tty_console_methods->own_selection_method = NULL;
+ tty_console_methods->get_foreign_selection_method = NULL;
+}
+
+#ifdef TIOCLINUX
+static Lisp_Object
+tty_get_foreign_selection (Lisp_Object selection_symbol, Lisp_Object target_type)
+{
+ /* This function can GC */
+ struct device *d = decode_device (Qnil);
+ int fd = DEVICE_INFD (d);
+ char c = 3;
+ Lisp_Object output_stream;
+ Lisp_Object terminal_stream;
+ Lisp_Object output_string;
+ struct gcpro gcpro1,gcpro2,gcpro3;
+
+ GCPRO3(output_stream,terminal_stream,output_string);
+
+ /* The ioctl() to paste actually puts things in the input queue of
+ ** the virtual console, so we need to trap that data, since we are
+ ** supposed to return the actual string selection from this
+ ** function.
+ */
+
+ /* I really hate doing this, but it doesn't seem to cause any
+ ** problems, and it makes the Lstream_read stuff further down
+ ** error out correctly instead of trying to indefinitely read from
+ ** the console.
+ **
+ ** There is no set_descriptor_blocking() function call, but in my
+ ** testing under linux, it has not proved fatal to leave the
+ ** descriptor in non-blocking mode.
+ **
+ ** William Perry Nov 5, 1999
+ */
+ set_descriptor_non_blocking (fd);
+
+ /* We need two streams, one for reading from the selected device,
+ ** and one to write the data into. There is no writable version
+ ** of the lisp-string lstream, so we make do with a resizing
+ ** buffer stream, and make a string out of it after we are
+ ** done.
+ */
+ output_stream = make_resizing_buffer_output_stream ();
+ terminal_stream = make_filedesc_input_stream (fd, 0, -1, LSTR_BLOCKED_OK);
+ output_string = Qnil;
+
+ /* We should arguably use a specbind() and an unwind routine here,
+ ** but I don't care that much right now.
+ ** ### FIXME?
+ */
+ if (NILP (output_stream) || NILP (terminal_stream))
+ {
+ /* Should we signal an error here? */
+ goto out;
+ }
+
+ if (ioctl (fd, TIOCLINUX, &c) < 0)
+ {
+ /* Could not get the selection - eek */
+ UNGCPRO;
+ return (Qnil);
+ }
+
+ while (1)
+ {
+ Bufbyte tempbuf[1024]; /* some random amount */
+ int i;
+ int size_in_bytes = Lstream_read (XLSTREAM (terminal_stream),
+ tempbuf, sizeof (tempbuf));
+
+ if (size_in_bytes <= 0)
+ {
+ /* end of the stream */
+ break;
+ }
+
+ /* convert CR->LF */
+ for (i = 0; i < size_in_bytes; i++)
+ {
+ if (tempbuf[i] == '\r')
+ {
+ tempbuf[i] = '\n';
+ }
+ }
+
+ Lstream_write (XLSTREAM (output_stream), tempbuf, size_in_bytes);
+ }
+
+ Lstream_flush (XLSTREAM (output_stream));
+
+ output_string = make_string (resizing_buffer_stream_ptr (XLSTREAM (output_stream)),
+ Lstream_byte_count (XLSTREAM (output_stream)));
+
+ Lstream_delete (XLSTREAM (output_stream));
+ Lstream_delete (XLSTREAM (terminal_stream));
+
+ out:
+ UNGCPRO;
+ return (output_string);
+}
+
+static Lisp_Object
+tty_selection_exists_p (Lisp_Object selection)
+{
+ return (Qt);
}
+#endif /* TIOCLINUX */
+#if 0
+static Lisp_Object
+tty_own_selection (Lisp_Object selection_name, Lisp_Object selection_value)
+{
+ /* There is no way to do this cleanly - the GPM selection
+ ** 'protocol' (actually the TIOCLINUX ioctl) requires a start and
+ ** end position on the _screen_, not a string to stick in there.
+ ** Lame.
+ **
+ ** William Perry Nov 4, 1999
+ */
+}
#endif
+
+/* This function appears to work once in a blue moon. I'm not sure
+** exactly why either. *sigh*
+** William Perry Nov 4, 1999
+*/
+static int
+tty_get_mouse_position (struct device *d, Lisp_Object *frame, int *x, int *y)
+{
+ Gpm_Event ev;
+ int num_buttons;
+
+ memset(&ev,'\0',sizeof(ev));
+ num_buttons = Gpm_GetSnapshot(&ev);
+
+ if (!num_buttons)
+ {
+ /* This means there are events pending... */
+ return(-1);
+ }
+ *x = ev.x;
+ *y = ev.y;
+ *frame = DEVICE_SELECTED_FRAME (d);
+ return (1);
+}
+
+static void
+tty_set_mouse_position (struct window *w, int x, int y)
+{
+ /* XXX
+ I couldn't find any GPM functions that set the mouse position.
+ Mr. Perry had left this function empty; that must be why.
+ karlheg
+ */
+}
+
+static int gpm_event_pending_p (int user_p)
+{
+ if (!NILP (gpm_event_queue))
+ {
+ /* We have a GPM event baked and ready to use */
+ return(1);
+ }
+ return (orig_event_pending_p (user_p));
+}
+
+static void gpm_next_event_cb (struct Lisp_Event *event)
+{
+ if (!NILP (gpm_event_queue))
+ {
+ Lisp_Object queued_event = dequeue_event (&gpm_event_queue, &gpm_event_queue_tail);
+ *event = *(XEVENT (queued_event));
+ return;
+ }
+
+ return (orig_next_event_cb (event));
+}
+
+static void hook_event_callbacks_once (void)
+{
+ static int hooker;
+
+ if (!hooker)
+ {
+ orig_event_pending_p = event_stream->event_pending_p;
+ orig_next_event_cb = event_stream->next_event_cb;
+ event_stream->event_pending_p = gpm_event_pending_p;
+ event_stream->next_event_cb = gpm_next_event_cb;
+ hooker = 1;
+ }
+}
+
+DEFUN ("gpm-enable", Fgpm_enable, 0, 1, 0, /*
+Toggle accepting of GPM mouse events.
+*/
+ (arg))
+{
+ Gpm_Connect conn;
+ int rval;
+ Lisp_Object gpm_process;
+ Lisp_Object gpm_filter;
+
+ if (NILP (arg))
+ {
+ turn_off_gpm ();
+ return (Qnil);
+ }
+
+ if (noninteractive)
+ {
+ error ("Can't connect to GPM in batch mode.");
+ }
+
+ hook_event_callbacks_once ();
+
+ /* DANGER DANGER.
+ ** Though shalt not call (gpm-enable t) after we have already
+ ** started, or stuff blows up.
+ */
+ if (!NILP (Fget_process ( build_string("gpm"))))
+ {
+ error ("GPM already enabled.");
+ }
+
+ /* Install the mouse position methods for the TTY console type */
+ CONSOLE_HAS_METHOD (tty, get_mouse_position);
+ CONSOLE_HAS_METHOD (tty, set_mouse_position);
+
+#ifdef TIOCLINUX
+ CONSOLE_HAS_METHOD (tty, get_foreign_selection);
+ CONSOLE_HAS_METHOD (tty, selection_exists_p);
+#if 0
+ CONSOLE_HAS_METHOD (tty, own_selection);
+#endif
+#endif /* TIOCLINUX */
+
+ conn.eventMask = GPM_DOWN|GPM_UP|GPM_MOVE|GPM_DRAG;
+ conn.defaultMask = GPM_MOVE;
+ conn.minMod = 0;
+ conn.maxMod = ((1<<KG_SHIFT)|(1<<KG_ALT)|(1<<KG_CTRL));
+
+ rval = Gpm_Open (&conn, 0);
+
+ switch (rval) {
+ case -1: /* General failure */
+ break;
+ case -2: /* We are running under an XTerm */
+ Gpm_Close();
+ break;
+ default:
+ /* Is this really necessary? */
+ set_descriptor_non_blocking (gpm_fd);
+ gpm_process = connect_to_file_descriptor (build_string ("gpm"), Qnil,
+ make_int (gpm_fd),
+ make_int (gpm_fd));
+
+ if (!NILP (gpm_process))
+ {
+ rval = 0;
+ Fprocess_kill_without_query (gpm_process, Qnil);
+ XSETSUBR (gpm_filter, &SFreceive_gpm_event);
+ set_process_filter (gpm_process, gpm_filter, 1);
+ }
+ else
+ {
+ Gpm_Close();
+ rval = -1;
+ }
+ }
+
+ return(rval ? Qnil : Qt);
+}
+
+void vars_of_gpmevent (void)
+{
+ gpm_event_queue = Qnil;
+ gpm_event_queue_tail = Qnil;
+ staticpro (&gpm_event_queue);
+ staticpro (&gpm_event_queue_tail);
+ pdump_wire (&gpm_event_queue);
+ pdump_wire (&gpm_event_queue_tail);
+}
+
+void syms_of_gpmevent (void)
+{
+ DEFSUBR (Freceive_gpm_event);
+ DEFSUBR (Fgpm_enable);
+}
+
+#endif /* HAVE_GPM */
Index: src/gpmevent.h
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/gpmevent.h,v
retrieving revision 1.3
diff -u -r1.3 gpmevent.h
--- gpmevent.h 1998/03/31 20:11:47 1.3
+++ gpmevent.h 1999/11/06 12:07:53
＠＠ -1,7 +0,0 ＠＠
-#ifndef _HAVE_GPM
-#define _HAVE_GPM
-
-int handle_gpm_read (struct Lisp_Event *event, struct console *con, int fd);
-void connect_to_gpm (struct console *con);
-
-#endif
Index: src/symsinit.h
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/symsinit.h,v
retrieving revision 1.31.2.13
diff -u -r1.31.2.13 symsinit.h
--- symsinit.h 1999/10/30 03:00:51 1.31.2.13
+++ symsinit.h 1999/11/06 12:07:54
＠＠ -148,6 +148,7 ＠＠
void syms_of_window (void);
void syms_of_xselect (void);
void syms_of_eldap (void);
+void syms_of_gpmevent (void);
/* Initialize the console types (dump-time but for reinit_). */
＠＠ -368,6 +369,7 ＠＠
void vars_of_xselect (void);
void reinit_vars_of_xselect (void);
void vars_of_eldap (void);
+void vars_of_gpmevent (void);
/* Initialize specifier variables (dump-time only). */
- 2 participants
- 1 comments
Re: Adaptive Fill Mode
25 years, 2 months
Hrvoje Niksic
Hrvoje Niksic
Gerald Gutierrez <gutier(a)intergate.bc.ca> writes:
> Under GNU Emacs, I can go into text-mode, enable auto-fill-mode, start
> typing garbage and get it to auto-indent subsequent lines:
>
> - asdf asdfklj asdkj asdklfj asdfklj fdasjkl asdlfkj asdfklj asdkj
> asdlfj asdlfkj asdkj asdf asdlf asdlfj asd asdfkl asdfkl asdfkj
> asdkj asdfkj asdklfj asdf
>
> How do I achieve this same effect in XEmacs?
M-x filladapt-mode RET
- 1 participants
- 0 comments
Adaptive Fill Mode
25 years, 2 months
Gerald Gutierrez
Gerald Gutierrez
Hi all.
Under GNU Emacs, I can go into text-mode, enable auto-fill-mode, start
typing garbage and get it to auto-indent subsequent lines:
- asdf asdfklj asdkj asdklfj asdfklj fdasjkl asdlfkj asdfklj asdkj
asdlfj asdlfkj asdkj asdf asdlf asdlfj asd asdfkl asdfkl asdfkj
asdkj asdfkj asdklfj asdf
How do I achieve this same effect in XEmacs?
I know that XEmacs has the same adaptive-fill-* variables, but 1) it
doesn't have the dash, and 2) even when I use one of the listed
characters (in adaptive-fill-regexp) as the start of the "paragraph",
it doesn't give me the leading spaces like GNU Emacs does:
> asdf asdlfkj asdfklja sdflkj asdfklj asdfklj asdfklj asdj asdfkl
asdfklj asdfj asdj asdklf asdklfj asdfklj asdfkla sdfklj asd asdfkj
asdfj asdfkla
If I hit M-q, I get:
> asdf asdlfkj asdfklja sdflkj asdfklj asdfklj asdfklj asdj asdfkl
> asdfklj asdfj asdj asdklf asdklfj asdfklj asdfkla sdfklj asd asdfkj
> asdfj asdfkla
which is obviously not what is wanted.
Thanks for any help.
- 1 participants
- 0 comments
Re: Give people with crashes more of a clue
25 years, 2 months
Hrvoje Niksic
Hrvoje Niksic
"Robin S. Socha" <robin(a)socha.net> writes:
> * Hrvoje Niksic <hniksic(a)iskon.hr> writes:
> > isn't it pointless to refer people to a PROBLEMS file that's not there
> > anymore? If we aren't already shipping PROBLEMS with XEmacs, maybe we
> > should start.
>
> Would it make any sense at all to not call the file PROBLEMS but
> PROBLEMS.<your_os>
Many items in PROBLEMS are OS-independent.
> or even README.<your_os>, and advertise it in the splash
> buffer/startup.el and the Help menu/menubar-items.el?
PROBLEMS is supposed to be primarily a helper file for people who
build XEmacs. Quite possibly some of the user-level documentation has
leaked there, but I don't think it's sufficient to warrant a place at
the menubar.
> I remember seeing intimidated users that only knew KDE. Yes, I know
> they should be using the RPMs.
Huh?
- 1 participants
- 0 comments