1 -*- coding: utf-8; mode: text; -*-
5 This branch implements support for opening multiple, different tty
6 devices and simultaneous X and tty frames from a single Emacs session.
9 Emacs is notoriously slow at startup, so most people use another
10 editor or emacsclient for quick editing jobs from the console.
11 Unfortunately, emacsclient was very awkward to use, because it did not
12 support opening a new Emacs frame on the current virtual console.
13 Now, with multi-tty support, it can do that. (Emacsclient starts up
16 Some Gnus users (including me) run Gnus in an X frame in its own Emacs
17 instance, which they typically leave running for weeks. It would be
18 nice if they could connect to this instance from a remote ssh session
19 and check their messages without opening a remote X frame or resorting
25 I'm Károly Lőrentey. My address: lorentey@elte.hu.
27 Comments, bug reports, suggestions and patches are welcome; send them
28 to multi-tty@lists.fnord.hu.
30 The following is a (sadly incomplete) list of people who have
31 contributed to the project by testing, submitting patches, bug
32 reports, and suggestions. Thanks!
34 Bernard Adrian <bernadrian@free.fr>
35 ARISAWA Akihiro <ari@mbf.ocn.ne.jp>
36 Vincent Bernat <bernat@luffy.cx>
37 Han Boetes <han@mijncomputer.nl>
38 Damien Cassou <damien.cassou@laposte.net>
39 Robert J. Chassell <bob@rattlesnake.com>
40 Romain Francoise <romain@orebokech.com>
41 Ami Fischman <ami@fischman.org>
42 Friedrich Delgado Friedrichs <friedel@nomaden.org>
43 IRIE Tetsuya <irie@t.email.ne.jp>
44 Yoshiaki Kasahara <kasahara@nc.kyushu-u.ac.jp>
45 Bas Kok <nekkobassu@yahoo.com>
46 Jurej Kubelka <Juraj.Kubelka@email.cz>
47 David Lichteblau <david@lichteblau.com>
48 Xavier Mallard <zedek@gnu-rox.org>
49 Istvan Marko <mi-mtty@kismala.com>
50 Ted Morse <morse@ciholas.com>
51 Gergely Nagy <algernon@debian.org>
52 Dan Nicolaescu <dann@ics.uci.edu>
53 Kalle Olavi Niemitalo <kon@iki.fi>
54 Mark Plaksin <happy@mcplaksin.org>
55 Francisco Borges <borges@let.rug.nl>
56 Frank Ruell <stoerte@dreamwarrior.net>
57 Dan Waber <dwaber@logolalia.com>
60 Richard Stallman was kind enough to review an earlier version of my
67 The multi-tty mailing list (discussion & bug reports):
69 Address: multi-tty@lists.fnord.hu
70 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty
71 Archive: http://news.gmane.org/gmane.emacs.multi-tty/
73 Commit notifications (read-only):
75 Address: multi-tty-commits@lists.fnord.hu
76 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty-commits
82 The branch is now very stable and almost full-featured. All of the
83 major problems have been fixed, only a few minor issues remain. (It
84 still needs to be ported to Windows/Mac/DOS, though.) Both multiple
85 tty device support and simultaneous X and tty frame support works
86 fine. Emacsclient has been extended to support opening new tty and X
87 frames. It has been changed open new Emacs frames by default.
89 The multi-tty branch has been scheduled for inclusion in the next
90 major release of Emacs (version 23). I expect the merge into the
91 development trunk to occur sometime during next year (2005), after the
92 merge of the Unicode branch.
94 Tested on GNU/Linux, Solaris 8, FreeBSD and OpenBSD. Please let me
95 know if you succeed or fail to use it on other platforms---I'll have a
96 few tricky test cases for you.
100 * GTK support. If you compile your Emacs with the GTK
101 toolkit, some functionality of multi-tty will be lost. In
102 particular, you will not be able to work on multiple X
103 displays at once. Current releases of GTK have limitations
104 and bugs that prevent full-blown multi-display support in
105 Emacs. (GTK crashes when Emacs tries to disconnect from an
106 X server.) Use the Lucid toolkit if you want to see a
107 complete feature set.
109 * The single-kboard mode.
111 If your multi-tty Emacs session seems to be frozen, you
112 probably have a recursive editing session or a pending
113 minibuffer prompt (which is a kind of recursive editing) on
114 another display. To unfreeze your session, switch to that
115 display and complete the recursive edit, for example by
116 pressing C-] (`abort-recursive-edit').
118 I am sorry to say that currently there is no way to break
119 out of this "single-kboard mode" from a frozen display. If
120 you are unable to switch to the display that locks the
121 others (for example because it is on a remote computer),
122 then you can use emacsclient to break out of all recursive
125 emacsclient -e '(top-level)'
127 Note that this (perhaps) unintuitive behaviour is by design.
128 Single-kboard mode is required because of an intrinsic Emacs
129 limitation that is very hard to eliminate. (This limitation
130 is related to the single-threaded nature of Emacs.)
132 I plan to implement better user notification and support for
133 breaking out of single-kboard mode from locked displays.
135 * Mac, Windows and DOS support is broken, doesn't even
136 compile. Multiple display support will probably not provide
137 new Emacs features on these systems, but the multi-tty
138 branch changed a few low-level interfaces, and the
139 system-dependent source files need to be adapted
140 accordingly. The changes are mostly trivial, so almost
141 anyone can help, if only by compiling the branch and
142 reporting the compiler errors. (It is not worth to do this
146 HOW TO GET THE BRANCH
147 ---------------------
149 The branch uses GNU Arch (http://www.gnuarch.org) for version control.
151 Retrieving the latest version of the branch:
153 tla register-archive -f http://lorentey.hu/arch/2004/
154 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
156 This incantation uses my private archive mirror that is hosted on a
157 relatively low-bandwidth site; if you are outside Hungary, you will
158 probably want to you use my secondary mirror: (Note that the -f option
159 will overwrite the archive location if you have previously registered
162 tla register-archive -f http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
163 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
165 http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
167 The Arch supermirror provides mirroring services for all public Arch
168 repositories. We have a mirror there, too, if you prefer.
170 tla register-archive -f http://mirrors.sourcecontrol.net/lorentey%40elte.hu--2004
171 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
173 My GPG key id is 0FB27A3F; it is available from
174 hkp://wwwkeys.eu.pgp.net/, or from my homepage at
175 http://lorentey.hu/rolam/gpg.html)
177 Don't worry if the above checkout takes a few minutes to complete;
178 once you have a source tree, updating it to the latest revision will
179 be _much_ faster. Use the following command for the update:
183 You can find more information about Arch on http://wiki.gnuarch.org/.
184 It's a wonderful source control system, I highly recommend it.
186 If you don't have tla, the branch has a homepage from which you can
187 download conventional patches against Emacs CVS HEAD:
189 http://lorentey.hu/project/emacs.html
194 If you run Debian, or a distribution based on Debian, you are welcome
195 to use our binary packages; put these lines in your /etc/apt/sources.list:
198 deb http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
199 deb-src http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
201 Note that these packages are intended solely to provide an easy way to
202 test the new multi-tty features. They are not to be taken as Emacs
203 releases, and it's a mistake to expect robust operation or any kind of
204 timely support for them. Do not install them, unless you'd like to
205 have your editor crash on you.
211 The multi-tty branch is compiled the same way as Emacs itself:
213 make maintainer-clean # (If you have compiled Emacs before)
215 ./configure --without-gtk <your favourite options>
219 If you have strange compilation errors, they may be caused by old
220 *.elc files that are left over from an earlier bootstrap. The `make
221 maintainer-clean' target deletes them, so it is a good idea to run
222 that before reporting a bug. (Emacs requires a clean recompilation
223 after certain kinds of source code changes.)
228 To test the multi-tty branch, start up the Emacs server with the
234 and then (from a shell prompt on another terminal) start emacsclient
236 emacsclient -t /optional/file/names... (for a tty frame)
237 emacsclient /optional/file/names... (for an X frame)
239 (Make sure both emacs and emacsclient are multi-tty versions.)
240 You'll hopefully have two fully working, independent frames on
241 separate terminals. The new frame is closed automatically when you
242 finish editing the specified files (C-x #), but delete-frame (C-x 5 0)
243 also works. Of course, you can create frames on more than two tty
246 Creating new frames on the same tty with C-x 5 2 works, and they
247 behave the same way as in previous Emacs versions. If you exit emacs,
248 all terminals should be restored to their previous states.
250 This is work in progress, and probably full of bugs. It is a good
251 idea to run emacs from gdb, so that you'll have a live instance to
252 debug if something goes wrong. Please send me your bug reports on our
253 mailing list: multi-tty@lists.fnord.hu
258 I think the best way to use the new Emacs is to have it running inside
259 a disconnected GNU screen session, and always use emacsclient for
260 normal work. One advantage of this is that not a single keystroke of
261 your work will be lost if the display device that you are using
262 crashes, or the network connection times out, or whatever. (I had an
263 extremely unstable X server for some time while I was developing these
264 patches, and running Emacs this way has saved me a number of M-x
265 recover-session invocations.)
267 I use the following two bash scripts to handle my Emacs sessions:
269 -------------------------------------------------------connect-emacs--
271 # Usage: connect-emacs <name> <args>...
273 # Connects to the Emacs instance called NAME. Starts up the instance
274 # if it is not already running. The rest of the arguments are passed
280 if [ -z "$name" ]; then
281 echo "Usage: connect_emacs <name> <args>..." >&2
284 preload-emacs "$name" wait
285 /usr/bin/emacsclient.emacs-multi-tty -s "$name" "$@"
286 ----------------------------------------------------------------------
288 -------------------------------------------------------preload-emacs--
290 # Usage: preload-emacs <name> [<waitp>]
292 # Preloads the Emacs instance called NAME in a detached screen
293 # session. Does nothing if the instance is already running. If WAITP
294 # is non-empty, the function waits until the server starts up and
295 # creates its socket; otherwise it returns immediately.
299 screendir="/var/run/screen/S-$USER"
300 serverdir="/tmp/emacs$UID"
301 emacs=/usr/bin/emacs-multi-tty # Or wherever you installed your multi-tty Emacs
303 if [ -z "$name" ]; then
304 echo "Usage: preload_emacs <name> [<waitp>]" >&2
308 if [ ! -e "$screendir"/*."$name" ]; then
309 if [ -e "$serverdir/$name" ]; then
310 # Delete leftover socket (for the wait option)
311 rm "$serverdir/$name"
313 screen -dmS "$name" "$emacs" -nw --eval "(setq server-name \"$name\")" -f server-start
315 if [ ! -z "$waitp" ]; then
316 while [ ! -e "$serverdir/$name" ]; do sleep 0.1; done
318 ----------------------------------------------------------------------
320 I have the following in my profile to have two instances automatically
321 preloaded for editing and email:
326 It is useful to set up short aliases for connect-emacs. I use the
329 alias edit="connect-emacs editor"
331 alias et="connect-emacs editor -t"
332 alias gnus="connect-emacs gnus"
344 For the NEWS file: (Needs much, much work)
346 ** Support for multiple terminal devices has been added.
348 *** You can specify a terminal device (`tty' parameter) and a terminal
349 type (`tty-type' parameter) to `make-terminal-frame'.
351 *** You can test for the presence of multiple terminal support by
352 testing for the `multi-tty' feature.
354 *** Emacsclient has been extended to support opening a new terminal
355 frame. Its behaviour has been changed to open a new Emacs frame by
356 default. Use the -c option to get the old behavior of opening
357 files in the currently selected Emacs frame.
359 *** A make-frame-on-tty function has been added to make it easier to
360 create frames on new terminals.
362 *** New functions: frame-tty-name, frame-tty-type, delete-tty,
363 terminal-local-value, set-terminal-local-value
365 terminal-id, terminal-parameters, terminal-parameter,
366 set-terminal-parameter
368 *** New variables: global-key-translation-map
370 *** The keymaps key-translation-map and function-key-map are now
373 ** Support for simultaneous graphical and terminal frames has been
376 *** The function `make-frame-on-display' now works during a terminal
377 session, and `make-frame-on-tty' works during a graphical session.
379 *** The `window-system' variable has been made frame-local.
381 *** The new `initial-window-system' variable contains the
382 `window-system' value for the first frame.
384 *** talk.el has been extended for multiple tty support.
389 (The rest of this file consists of my development notes and as such it
390 is probably not very interesting for anyone else.)
395 ** Report GTK multi-display problems to GTK maintainers. For extra
398 Currently you can not connect to new X displays when you compile
399 Emacs with GTK support. If you want to play around with GTK
400 multidisplay (and don't mind core dumps), you can edit src/config.h
401 and define HAVE_GTK_MULTIDISPLAY there by hand.
403 ** Audit `face-valid-attribute-values' usage in customize and
404 elsewhere. Its return value depends on the current window system.
405 Replace static initializers using it with runtime functions. For
406 example, custom's buttons are broken on non-initial device types.
408 ** Possibly turn off the double C-g feature when there is an X frame.
409 C.f. (emacs)Emergency Escape.
411 ** frames-on-display-list should also accept frames.
413 ** I smell something funny around pop_kboard's "deleted kboard" case.
414 Determine what are the circumstances of this case, and fix any
415 bug that comes to light.
417 ** Consider the `tty-type' frame parameter and the `display-tty-type'
418 function. They serve the exact same purpose. I think it may be
419 a good idea to eliminate one of them, preferably `tty-type'.
421 ** The handling of lisp/term/*.el, and frame creation in general, is a
422 big, big mess. How come the terminal-specific file is loaded by
423 tty-create-frame-with-faces? I don't think it is necessary to load
424 these files for each frame; once per terminal should be enough.
425 Update: lisp/term/*.el is not loaded repeatedly anymore, but
426 faces.el still needs to be cleaned up.
428 ** Fix frame-set-background-mode in this branch. It was recently
429 changed in CVS, and frame.el in multi-tty has not yet been adapted
430 for the changes. (It needs to look at
431 default-frame-background-mode.) (Update: maybe it is fixed now;
434 ** I think `(set-)terminal-local-value' and the terminal parameter
435 mechanism should be integrated into a single framework.
437 ** Add the following hooks: after-delete-frame-hook (for server.el,
438 instead of delete-frame-functions),
439 after-delete-terminal-functions, after-create-terminal-functions.
441 ** Fix set-input-mode for multi-tty. It's a truly horrible interface;
442 what if we'd blow it up into several separate functions (with a
443 compatibility definition)?
445 ** BULK RENAME: The `display-' prefix of new Lisp-level functions
446 conflicts with stuff like `display-time-mode'. Use `device-'
447 or `terminal-' instead. I think I prefer `terminal-'.
449 It turns out that most of the offending Lisp functions were defined
450 in the trunk. Therefore, compatibility aliases should be defined
451 for the following names:
453 display-color-cells terminal-color-cells
454 display-color-p terminal-color-p
455 display-graphic-p terminal-graphic-p
456 display-grayscale-p terminal-grayscale-p
457 display-images-p terminal-images-p
458 display-mm-height terminal-mm-height
459 display-mm-width terminal-mm-width
460 display-mouse-p terminal-mouse-p
461 display-multi-font-p terminal-multi-font-p
462 display-multi-frame-p terminal-multi-frame-p
463 display-pixel-height terminal-pixel-height
464 display-pixel-width terminal-pixel-width
465 display-pixels-per-inch terminal-pixels-per-inch
466 display-planes terminal-planes
467 display-popup-menus-p terminal-popup-menus-p
468 display-save-under terminal-save-under
469 display-screens terminal-screens
470 display-supports-face-attributes-p terminal-supports-face-attributes-p
471 display-visual-class terminal-visual-class
472 framep-on-display framep-on-terminal
473 frames-on-display-list frames-on-terminal-list
475 The following functions were introduced in the multi-tty branch, and
476 can be renamed without aliases:
478 display-controlling-tty-p terminal-controlling-tty-p
479 display-list terminal-list
480 display-live-p terminal-live-p
481 display-name terminal-name
482 display-tty-type terminal-tty-type
483 frame-display terminal-of-frame
484 delete-display delete-terminal
486 ** The single-keyboard mode of MULTI_KBOARD is extremely confusing
487 sometimes; Emacs does not respond to stimuli from other keyboards.
488 At least a beep or a message would be important, if the single-mode
489 is still required to prevent interference. (Reported by Dan
492 Update: selecting a region with the mouse enables single_kboard
493 under X. This is very confusing.
495 Update: After discussions with Richard Stallman, this will be
496 resolved by having locked displays warn the user to wait, and
497 introducing a complex protocol to remotely bail out of
498 single-kboard mode by pressing C-g.
500 Update: Warning the user is not trivial to implement, as Emacs has
501 only one echo area, shared by all frames. Ideally the warning
502 should not be displayed on the display that is locking the others.
503 Perhaps the high probability of user confusion caused by
504 single_kboard mode deserves a special case in the display code.
505 Alternatively, it might be good enough to signal single_kboard mode
506 by changing the modelines or some other frame-local display element
507 on the locked out displays.
509 ** The session management module is prone to crashes when the X
510 connection is closed and then later I try to connect to a new X
513 #0 0xb7ebc806 in SmcGetIceConnection () from /usr/X11R6/lib/libSM.so.6
514 #1 0x080e6641 in x_session_check_input (bufp=0xbf86c9c0) at xsmfns.c:144
515 #2 0x080d3bbc in XTread_socket (device=0xa722ff8, expected=1, hold_quit=0xbf86ca90) at xterm.c:7037
516 #3 0x080fa404 in read_avail_input (expected=1) at keyboard.c:6696
517 #4 0x080fa4ca in handle_async_input () at keyboard.c:6900
518 #5 0x080d51fa in x_term_init (display_name=162628899, xrm_option=0x0, resource_name=0x857068c "emacs") at xterm.c:10622
519 #6 0x080d920e in x_display_info_for_name (name=162628899) at xfns.c:3975
520 #7 0x080d92f9 in check_x_display_info (object=1) at xfns.c:274
521 #8 0x080d97b8 in Fx_create_frame (parms=151221485) at xfns.c:3016
522 #9 0x0815bf72 in Ffuncall (nargs=2, args=0xbf86ceec) at eval.c:2851
524 I installed a workaround to prevent this. The X session manager is
525 only contacted when the very first display in the Emacs session is
526 an X display. Also, x_delete_display() on this display aborts
527 session management, and XTread_socket only calls
528 x_session_check_input when it is called for the display that the
529 session was opened on. While this does not really fix the bug, it
530 makes it much less frequent, because session manager support will
531 not normally be enabled when Emacs can survive the shutdown of the
534 See if xsmfns.c should be updated.
536 ** Hunt down display-related functions in frame.el and extend them all
537 to accept display ids.
539 ** rif->flush_display_optional (NULL) calls should be replaced by a
542 ** The set-locale-environment hack (adding the DISPLAY option) should
543 be replaced with a clean design.
545 ** standard-display-table should be display-local.
546 standard-display-european should be display-local.
548 ** With iswitchb-default-method set to 'always-frame, only frames on
549 the current display should be considered. This might involve
550 extending `get-buffer-window'.
552 ** Have a look at Vlocale_coding_system. Seems like it would be a
553 tedious job to localize it, although most references use it for
554 interfacing with libc and are therefore OK with the global
557 Exceptions found so far: x-select-text and
558 x-cut-buffer-or-selection-value.
560 ** Have a look at fatal_error_hook.
562 ** Have a look at set_frame_matrix_frame.
564 ** Check if we got term-setup-hook right.
566 ** I think tip_frame should be display-local.
568 ** Check display reference count handling in x_create_tip_frame.
570 ** make-frame does not correctly handle extra parameters in its
573 (frame-parameter (make-frame (list (cons 'foobar 42))) 'foobar)
576 (This is likely an error in the CVS trunk.)
578 ** Dan Nicolaescu suggests that -nw should be added as an alias for -t
579 in emacsclient. Good idea. (Alas, implementing this is not
580 trivial, getopt_long does not seem to support two-letter ``short''
581 options. Patches are welcome.)
583 ** Mark Plaksin suggests that emacsclient should accept the same
584 X-related command-line arguments as Emacs. Most of the X-related
585 argument-handling is done in Lisp, so this should be quite easy to
588 ** Gergely Nagy suggests that C-x # should only kill the current
589 frame, not any other emacsclient frame that may have the same file
590 opened for editing. I think I agree with him.
592 ** Very strange bug: visible-bell does not work on secondary
593 terminals in xterm and konsole. The screen does flicker a bit,
594 but it's so quick it isn't noticable.
596 ** Move baud_rate to struct display.
598 ** Implement support for starting an interactive Emacs session without
599 an initial frame. (The user would connect to it and open frames
600 later, with emacsclient.)
602 ** Fix Mac support (I can't do this entirely myself). Note that the
603 current state of Mac-specific source files in the multi-tty tree
604 are not useful; before starting work on Mac support, revert to
605 pristine, pre-multi-tty versions.
607 ** Fix W32 support (I can't do this entirely myself). Note that the
608 current state of W32-specific source files in the multi-tty tree
609 are not useful; before starting work on W32 support, revert to
610 pristine, pre-multi-tty versions.
612 ** Fix DOS support (I can't do this entirely myself). Note that the
613 current state of DOS-specific source files in the multi-tty tree
614 are not useful; before starting work on DOS support, revert to
615 pristine, pre-multi-tty versions.
617 ** Do a grep on XXX and ?? for more issues.
619 ** Understand Emacs's low-level input system (it's black magic) :-)
620 What exactly does interrupt_input do? I tried to disable it for
621 raw secondary tty support, but it does not seem to do anything
622 useful. (Update: Look again. X unconditionally enables this, maybe
623 that's why raw terminal support is broken again. I really do need
624 to understand input.)
625 (Update: I am starting to understand the read_key_sequence->read-char
626 ->kbd_buffer_get_event->read_avail_input->read_socket_hook path. Yay!)
628 ** flow-ctrl.el must be updated.
630 ** Fix stuff_char for multi-tty. Doesn't seem to be of high priority.
635 (ex-TODO items with explanations.)
637 -- Introduce a new struct for terminal devices.
639 (Done, see struct tty_output. The list of members is not yet
642 -- Change the bootstrap procedure to initialize tty_list.
644 (Done, but needs review.)
646 -- Change make-terminal-frame to support specifying another tty.
648 (Done, new frame parameters: `tty' and `tty-type'.)
650 -- Implement support for reading from multiple terminals.
652 (Done, read_avail_input tries to read from each terminal, until one
653 succeeds. MULTI_KBOARD is not used. Secondary terminals don't send
656 (Update: They do, now.)
658 (Update2: After enabling X, they don't.)
660 -- other-frame should cycle through the frames on the `current'
663 (Done, by trivially modifiying next_frame and prev_frame.)
665 -- Support different terminal sizes.
669 -- Make sure terminal resizes are handled gracefully. (Could be
672 (Done. We don't get automatic SIGWINCH for additional ttys,
675 -- Extend emacsclient to automatically open a new tty when it connects
678 (Done. It's an ugly hack, needs more work.)
680 -- Redisplay must refresh the topmost frame on *all* terminals, not
681 just the initial terminal.
683 (Done, but introduced an ugly redisplay problems. Ugh.)
685 -- Fix redisplay problems.
687 (Done; it turned out that the entire Wcm structure must be moved
688 inside tty_output. Why didn't I catch this earlier?)
690 -- Provide a way for emacsclient to tell Emacs that the tty has been
693 (Done, simply forward the SIGWINCH signal.)
695 -- Each keypress should automatically select the frame corresponding
696 to the terminal that it was coming from. This means that Emacs
697 must know from which terminal the last keyboard event came from.
699 (Done, it was quite simple, the input event system already
700 supported multiple frames.)
702 -- Fix SIGIO issue with secondary terminals.
704 (Done, emacsclient signals Emacs after writing to the proxy pseudo
705 terminal. Note that this means that multi-tty does not work with
708 (Update: This is bullshit. There is a read_input_waiting function,
709 extend that somehow.)
711 (Update of update: The first update was not right either, extending
712 read_input_waiting was not necessary. Secondary ttys do seem to
713 send signals on input.)
715 (Update^3: Not any more.)
717 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
718 parameters from the currently selected terminal before the global
723 -- Put all cached terminal escape sequences into struct tty_output.
724 Currently, they are still stored in global variables, so we don't
725 really support multiple terminal types.
727 (Done. It was not fun.)
729 -- Implement sane error handling after initialization. (Currently
730 emacs exits if you specify a bad terminal type.) The helpful error
731 messages must still be provided when Emacs starts.
735 -- Implement terminal deletion, i.e., deleting local frames, closing
736 the tty device and restoring its previous state without exiting
739 (Done, but at the moment only called when an error happens during
740 initialization. There is a memory corruption error around this
741 somewhere.) (Update: now it is fully enabled.)
743 -- Implement automatic deletion of terminals when the last frame on
744 that terminal is closed.
748 -- Restore tty screen after closing the terminal.
750 (Done, we do the same as Emacs 21.2 for all terminals.)
752 -- 'TERM=dumb src/emacs' does not restore the terminal state.
756 -- C-g should work on secondary terminals.
758 (Done, but the binding is not configurable.)
760 -- Deal with SIGHUP in Emacs and in emacsclient. (After this, the
761 server-frames may be removed from server.el.)
763 (Done, nothing to do. It seems that Emacs does not receive SIGHUP
764 from secondary ttys, which is actually a good thing.) (Update: I
765 think it would be a bad idea to remove server-frames.)
767 -- Change emacsclient/server.el to support the -t argument better,
768 i.e. automatically close the socket when the frame is closed.
770 (Seems to be working OK.)
772 -- Fix mysterious memory corruption error with tty deletion. To
773 trigger it, try the following shell command:
775 while true; do TERM=no-such-terminal-definition emacsclient -h; done
777 Emacs usually dumps core after a few dozen iterations. (The bug
778 seems to be related to the xfreeing or bzeroing of
779 tty_output.Wcm. Maybe there are outside references to struct Wcm?
780 Why were these vars collected into a struct before multi-tty
783 (Done. Whew. It turned out that the problem had nothing to do
784 with hypothetical external references to Wcm, or any other
785 tty_output component; it was simply that delete_tty closed the
786 filehandles of secondary ttys twice, resulting in fclose doubly
787 freeing memory. Utterly trivial matter. I love the C's memory
788 management, it puts hair on your chest.)
790 -- Support raw secondary terminals. (Note that SIGIO works only on
791 the controlling terminal.) Hint: extend read_input_waiting for
792 multiple ttys and hopefully this will be fixed.
794 (Done, it seems to have been working already for some time. It
795 seems F_SETOWN does work, after all. Not sure what made it fail
796 earlier, but it seems to be fixed (there were several changes
797 around request_sigio, maybe one of them did it).
798 read_input_waiting is only used in sys_select, don't change
799 it.) (Update: After adding X support, it's broken again.)
800 (Update^2: No it isn't.) :-)
802 -- Find out why does Emacs abort when it wants to close its
803 controlling tty. Hint: chan_process[] array. Hey, maybe
804 noninterrupt-IO would work, too? Update: no, there is no process
807 (Done. Added add/delete_keyboard_wait_descriptor to
808 term_init/delete_tty. The hint was right, in a way.)
810 -- Issue with SIGIO: it needs to be disabled during redisplay. See if
811 fcntl kernel behaviour could be emulated by emacsclient.
813 (Done. Simply disabled the SIGIO emulation hack in emacsclient.)
814 (Update: it was added back.) (Update^2: and removed again.)
816 -- server.el: There are issues with saving files in buffers of closed
817 clients. Try editing a file with emacsclient -f, and (without
818 saving it) do a delete-frame. The frame is closed without
819 question, and a surprising confirmation prompt appears in another
822 (Done. delete-frame now asks for confirmation if it still has
823 pending buffers, and modified buffers don't seem to be deleted.)
825 -- emacsclient.el, server.el: Handle eval or file open errors when
830 -- Make parts of struct tty_output accessible from Lisp. The device
831 name and the type is sufficient.
833 (Done, see frame-tty-name and frame-tty-type.)
835 -- Export delete_tty to the Lisp environment, for emacsclient.
837 (Done, see delete-tty.)
839 -- Get rid of the accessor macros in termchar.h, or define macros for
844 -- Move device-specific parameters (like costs) commonly used by
845 device backends to a common, device-dependent structure.
847 (Done. See struct display_method in termhooks.h.)
851 (Done. Well, it seems to be working.)
853 -- Allow simultaneous X and tty frames. (Handling input could be
854 tricky. Or maybe not.)
856 (Done. Allowed, that is. It is currently extremely unstable, to
857 the point of being unusable. The rif variable causes constant
858 core dumps. Handling input is indeed tricky.)
860 -- Rewrite multi-tty input in terms of MULTI_KBOARD.
862 (Done. In fact, there was no need to rewrite anything, I just
863 added a kboard member to tty_display_info, and initialized the
864 frame's kboard from there.)
866 -- Fix rif issue with X-tty combo sessions. IMHO the best thing to do
867 is to get rid of that global variable (and use the value value in
868 display_method, which is guaranteed to be correct).
870 (Done, did exactly that. Core dumps during combo sessions became
871 much rarer. In fact, I have not yet met a single one.)
873 -- Add multi-tty support to talk.el.
877 -- Clean up the source of emacsclient. It is a mess.
879 (Done, eliminated stupid proxy-pty kludge.)
881 -- Fix faces on tty frames during X-tty combo sessions. There is an
882 init_frame_faces call in init_sys_modes, see if there is a problem
885 (Done, there was a stupid mistake in
886 Ftty_supports_face_attributes_p. Colors are broken, though.)
888 -- C-x 5 2, C-x 5 o, C-x 5 0 on an emacsclient frame unexpectedly
889 exits emacsclient. This is a result of trying to be clever with
890 delete-frame-functions.
892 (Fixed, added delete-tty-after-functions, and changed server.el to
895 -- Something with (maybe) multi-keyboard support broke function keys
896 and arrows on ttys during X+tty combo sessions. Debug this.
898 (I can't reproduce it, maybe the terminal type was wrong.)
900 -- Fix input from raw ttys (again).
902 (Now it seems to work all right.)
904 -- During an X-tty combo session, a (message "Hello") from a tty frame
905 goes to the X frame. Fix this.
907 (Done. There was a safeguard against writing to the initial
908 terminal frame during bootstrap which prevented echo_area_display
909 from working correctly on a tty frame during a combo session.)
911 -- If there are no frames on its controlling terminal, Emacs should
912 exit if the user presses C-c there.
914 (Done, as far as possible. See the SIGTERM comment in
915 interrupt_signal on why this seems to be impossible to solve this
918 -- During an X session, Emacs seems to read from stdin. Also, Emacs
919 fails to start without a controlling tty.
921 (Fixed by replacing the troublesome termcap display with a dummy
922 bootstrap display during bootstrap.
924 -- Do tty output through struct display, like graphical display
929 -- Define an output_initial value for output_method for the initial
930 frame that is dumped with Emacs. Checking for this frame (e.g. in
931 cmd_error_internal) is ugly.
933 (Done, breaking interactive temacs.)
935 -- The command `emacsclient -t -e '(delete-frame)'' fails to exit.
939 -- frame-creation-function should always create a frame that is on the
940 same display as the selected frame. Maybe frame-creation-function
941 should simply be removed and make-frame changed to do the right
944 (Done, with a nice hack. frame-creation-function is now frame-local.)
946 -- Fix C-g on raw ttys.
948 (Done. I disabled the interrupt/quit keys on all secondary
949 terminals, so Emacs sees C-g as normal input. This looks like an
950 overkill, because emacsclient has extra code to pass SIGINT to
951 Emacs, so C-g should remain the interrupt/quit key on emacsclient
952 frames. See the next entry why implementing this distinction would
955 -- Make sure C-g goes to the right frame with ttys. This is hard, as
956 SIGINT doesn't have a tty parameter. :-(
958 (Done, the previous change fixes this as a pleasant side effect.)
960 -- I have seen a case when Emacs with multiple ttys fell into a loop
961 eating 100% of CPU time. Strace showed this loop:
964 kill(30284, SIGIO) = 0
965 --- SIGIO (I/O possible) @ 0 (0) ---
966 ioctl(6, FIONREAD, [0]) = -1 EIO (Input/output error)
967 ioctl(5, FIONREAD, [0]) = -1 EIO (Input/output error)
968 ioctl(0, FIONREAD, [0]) = 0
969 sigreturn() = ? (mask now [])
970 gettimeofday({1072842297, 747760}, NULL) = 0
971 gettimeofday({1072842297, 747806}, NULL) = 0
972 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
973 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
974 gettimeofday({1072842297, 748245}, NULL) = 0
976 I have seen something similar with a single X frame, but have not
977 been able to reproduce it for debugging.
979 Update: This may have been caused by checking for nread != 0
980 instead of nread > 0 after calling read_socket_hook in
983 (Fixed. This was caused by unconditionally including stdin in
984 input_wait_mask in init_process. The select call in
985 wait_reading_process_input always returned immediately, indicating
986 that there is pending input from stdin, which nobody read.
988 Note that the above strace output seems to be an unrelated but
989 similar bug. I think that is now fixed.)
991 -- Exiting Emacs while there are emacsclient frames doesn't restore the
992 ttys to their default states.
994 (This seems to be fixed by some previous change.)
996 -- Allow opening an X session after -nw.
1000 -- Fix color handling during tty+X combo sessions. (It seems that tty
1001 sessions automatically convert the face colors to terminal colors
1002 when the face is loaded. This conversion must happen instead on
1003 the fly in write_glyphs, which might be problematic, as color
1004 approximation is currently done in lisp (term/tty-colors.el).)
1005 (Update: hm, colors seem to work fine if I start emacs with -nw and
1006 then create an X frame. Maybe it's just a small buglet somewhere.)
1008 (Seems to be fixed. The problem was in startup.el, it did not
1009 initialize tty colors when the initial window system was
1012 -- emacs -nw --eval '(y-or-n-p "Foobar")' segfaults. (Reported by
1015 (Fixed, there was a keyboard initialization problem.)
1017 -- Fix interactive use of temacs. There are face-related SEGVs, most
1018 likely because of changes in realize_default_face, realize_face.
1022 -- Don't exit Emacs when the last X connection fails during a
1023 multi-display session.
1027 -- Dan Nicolaescu noticed that starting emacsclient on the same
1028 terminal device that is the controlling tty of the Emacs process
1029 gives unexpected results.
1033 -- Istvan Marko reported that Emacs hang on ttys if it was started
1034 from a shell script.
1036 (Fixed. There was a bug in the multi-tty version of
1037 narrow_foreground_group. tcsetpgrp blocks if it is called from a
1038 process that is not in the same process group as the tty.)
1040 -- emacsclient -t from an Emacs term buffer does not work, complains
1041 about face problems. This can even lock up Emacs (if the recursive
1042 frame sets single_kboard). Update: the face problems are caused by
1043 bugs in term.el, not in multi-tty. The lockup is caused by
1044 single_kboard mode, and is not easily resolvable. The best thing to
1045 do is to simply refuse to create a tty frame of type `eterm'.
1047 (Fixed, changed emacsclient to check for TERM=eterm. The face
1048 complaints seem to be caused by bugs in term.el; they are not
1049 related to multi-tty.)
1051 -- Find out the best way to support suspending Emacs with multiple
1052 ttys. My guess: disable it on the controlling tty, but from other
1053 ttys pass it on to emacsclient somehow. (It is (I hope) trivial to
1054 extend emacsclient to handle suspend/resume. A `kill -STOP' almost
1057 (Done. I needed to play with signal handling and the server
1058 protocol a bit to make emacsclient behave as a normal UNIX program
1059 wrt foreground/background process groups.)
1061 -- There is a flicker during the startup of `emacs -nw'; it's as if
1062 the terminal is initialized, reset and then initialialized again.
1063 Debug this. (Hint: narrow_foreground_group is called twice during
1068 -- Robert Chassell has found serious copy-paste bugs with the
1069 multi-tty branch. There seem to be redisplay bugs while copying
1070 from X to a terminal frame. Copying accented characters do not
1073 (Patch-124 should fix this, by changing the interprogram-*-function
1074 variables to be frame-local, as suggested by Mark Plaksin
1075 (thanks!). I think that the redisplay bugs are in fact not bugs,
1076 but delays caused by single_kboard --> perhaps MULTI_KBOARD should
1079 -- frame-creation-function was removed, which might be a bad idea.
1080 Think up a compatible solution.
1082 (It was an internal interface that may be changed when necessary.)
1084 -- Change Lisp code not to (getenv "TERM"); use the `tty-type' frame
1085 parameter or the frame-tty-type function instead. (M-x tags-search
1086 "TERM" helps with this.) Update: Actually, all getenv invocations
1087 should be checked for multi-tty compatibility, and an interface
1088 must be implemented to get the remote client's environment.
1090 (Done. Only getenv calls in lisp/term/*.el were changed; other
1091 calls should be mostly left as they are.)
1093 -- Add an elaborate mechanism for display-local variables. (There are
1094 already a few of these; search for `terminal-local' in the Elisp
1097 (Not needed. Display-local variables could be emulated by
1098 frame-local variables.)
1100 -- Emacs assumes that all terminal frames have the same locale
1101 settings as Emacs itself. This may lead to bogus results in a
1102 multi-locale setup. (E.g., while logging in from a remote client
1103 with a different locale.)
1104 (Update after new bugreport by Friedrich Delgado Friedrichs:
1105 (at least) the structs terminal_coding and keyboard_coding in
1106 coding.c must be moved to struct display, and the Lisp interface
1107 [set-]keyboard-coding-system must be adapted for the change.)
1109 (Fixed. Emacs now uses the locale settings as seen by the
1110 emacsclient process for server tty frames.)
1111 (Update: Not really; Vlocale_coding_system is still global.)
1113 -- Make `struct display' accessible to Lisp programs. Accessor functions:
1115 (displayp OBJECT): Returns t if OBJECT is a display.
1116 => Implemented as display-live-p.
1118 (display-list): Returns list of currently active displays.
1121 (selected-display): Returns the display object of the selected frame.
1122 => Not strictly necessary, but implemented anyway.
1124 (frame-display FRAME): Returns the display object of FRAME.
1127 (display-frames DISPLAY): Returns a list of frames on DISPLAY.
1128 => Already implemented, see frames-on-display-list.
1130 (display-type DISPLAY): Returns the type of DISPLAY, as a
1131 symbol. (See `framep'.)
1132 => Implemented as display-live-p.
1134 (display-device DISPLAY): Returns the name of the device that
1135 DISPLAY uses, as a string. (E.g: "/dev/pts/16", or
1137 => Implemented as display-name.
1141 See next issue why this is necessary.
1143 (Update: The consensus on emacs-devel seems to be to do this via
1144 integer identifiers. That's fine by me.)
1148 -- The following needs to be supported:
1155 The cleanest way to solve this is to allow multiple displays on the
1156 same terminal device; each new emacsclient process should create
1157 its own display. As displays are currently identified by their
1158 device names, this is not possible until struct display becomes
1159 accessible as a Lisp-level object.
1163 -- Miles Bader suggests that C-x C-c on an emacsclient frame should
1164 only close the frame, not exit the entire Emacs session. Update:
1165 see above for a function that does this. Maybe this should be the
1168 (Done. This is the new default. No complaints so far.)
1170 -- Clean up the frame-local variable system. I think it's ugly and
1171 error-prone. But maybe I just haven't yet fully understood it.
1173 (Nothing to do. It doesn't seem ugly any more. It's rather clever.)
1175 -- Support multiple character locales. A version of
1176 `set-locale-environment' needs to be written for setting up
1177 display-local settings on ttys. I think calling
1178 set-display-table-and-terminal-coding-system and
1179 set-keyboard-coding-system would be enough. The language
1180 environment itself should remain a global setting.
1182 (Done, by an ugly hack.)
1184 -- The terminal customization files in term/*.el tend to change global
1185 parameters, which may confuse Emacs with multiple displays. Change
1186 them to tweak only frame-local settings, if possible. (They tend
1187 to call define-key to set function key sequences a lot.)
1189 (Done, by making `function-key-map' terminal-local (i.e., part of
1190 struct kboard). This has probably covered all the remaining problems.)
1192 -- Make `function-key-map' and `key-translation-map' terminal-local.
1196 -- Implement `terminal-local-value' and `set-terminal-local-value' to
1197 allow deterministic access to terminal local bindings. The
1198 encode-kb package can not set up `key-translation-map' without
1199 these. The terminal-local bindings seem to be independent of what
1204 -- xt-mouse.el needs to be adapted for multi-tty. It currently
1205 signals an error on kill-emacs under X, which prevents the user
1206 from exiting Emacs. (Reported by Mnemonikk on freenode.)
1211 -- Having {reset,init}_all_sys_modes in set-input-mode breaks arrow
1212 keys on non-selected terminals under screen, and sometimes on other
1213 terminal types as well. The other function keys continue to work
1214 fine. Sometimes faces on these screens become garbled.
1216 This only seems to affect displays that are of the same terminfo
1217 type as the selected one. Interestingly, in screen Emacs normally
1218 reports the up arrow key as `M-o A', but after the above SNAFU, it
1219 complains about `M-[ a'. UNIX ttys are a complete mystery to me,
1220 but it seems the reset-reinitialize cycle somehow leaves the
1221 non-selected terminals in a different state than usual. I have no
1222 idea how this could happen.
1224 Currently set-input-mode resets only the currently selected
1225 terminal, which seems to somehow work around the problem.
1229 Dan Nicolaescu <dann@ics.uci.edu> writes:
1230 > Some terminals have 2 modes for cursor keys: Application Mode where
1231 > the cursor keys transmit the codes defined in the terminfo entry, and
1232 > Cursor mode. Applications have to send the smkx and rmkx terminfo
1233 > strings to switch between the 2 modes. So Emacs (and emacsclient) have
1234 > to send smkx when initializing and rmkx when quitting (or on
1237 (I think patch-370 fixed this.)
1239 -- This long-standing bug (first reported by Han Boetes) seems to come
1240 and go all the time. It is time to track it down and fix it.
1245 # From another xterm:
1246 emacsclient -e '(y-or-n-p "Do you want me to crash? ")'
1247 # Notice how the answer ends up in the *scratch* buffer
1251 (Fixed in patch-414 after detailed analysis by Kalle Olavi Niemitalo.)
1253 -- normal-erase-is-backspace-mode in simple.el needs to be updated for
1254 multi-tty (rep. by Dan Waber). (The Delete key is broken on X
1257 (Fixed in patch-427.)
1259 -- I think keyboard-translate-table should be made terminal-local.
1261 (Done in patch-431.)
1263 -- The semantics of terminal-local variables are confusing; it is not
1264 clear what binding is in effect in any given time. See if
1265 current_kboard (or at least the terminal-local bindings exported to
1266 Lisp) might be changed to be tied to the selected frame instead.
1267 Currently, `function-key-map' and `key-translation-map' may be
1268 accessed reliably only using the hackish
1269 `(set-)terminal-local-value' functions.
1271 Perhaps there should be a difference between `last-command' &co.
1272 and these more conventional configuration variables.
1273 (E.g. `symbol-value' would use current_kboard to access
1274 `last-command', but SELECTED_FRAME()->display->kboard to get the
1275 value of `function-key-map'.
1277 (Fixed in patch-434.)
1279 -- If the first key pressed on a new tty terminal is a function key,
1280 it is not recognized correctly. May be related to the bug below.
1282 (Seems to have been fixed as a side effect of patch-434. "The bug
1283 below" was the set-input-mode madness.)
1285 (Update: this bug was fixed for good in patch-449. It was tracked
1286 down to a bug in `read_key_sequence': it failed to reinitialize its
1287 local function-key-map/key-translation-map references when it
1288 switched keyboards. I don't understand why did this bug only
1289 appear on brand new frames, though!)
1291 -- Disable connecting to a new X display when we use the GTK toolkit.
1293 (Disabled in patch-450.)
1295 ;;; arch-tag: 8da1619e-2e79-41a8-9ac9-a0485daad17d