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 Francisco Borges <borges@let.rug.nl>
39 Damien Cassou <damien.cassou@laposte.net>
40 Robert J. Chassell <bob@rattlesnake.com>
41 Romain Francoise <romain@orebokech.com>
42 Ami Fischman <ami@fischman.org>
43 Friedrich Delgado Friedrichs <friedel@nomaden.org>
44 IRIE Tetsuya <irie@t.email.ne.jp>
45 Yoshiaki Kasahara <kasahara@nc.kyushu-u.ac.jp>
46 Bas Kok <nekkobassu@yahoo.com>
47 Jurej Kubelka <Juraj.Kubelka@email.cz>
48 David Lichteblau <david@lichteblau.com>
49 Xavier Mallard <zedek@gnu-rox.org>
50 Istvan Marko <mi-mtty@kismala.com>
51 Ted Morse <morse@ciholas.com>
52 Gergely Nagy <algernon@debian.org>
53 Dan Nicolaescu <dann@ics.uci.edu>
54 Kalle Olavi Niemitalo <kon@iki.fi>
55 Mark Plaksin <happy@mcplaksin.org>
56 Frank Ruell <stoerte@dreamwarrior.net>
57 Tom Schutzer-Weissmann <trmsw@yahoo.co.uk>
58 Dan Waber <dwaber@logolalia.com>
61 Richard Stallman was kind enough to review an earlier version of my
68 The multi-tty mailing list (discussion & bug reports):
70 Address: multi-tty@lists.fnord.hu
71 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty
72 Archive: http://news.gmane.org/gmane.emacs.multi-tty/
74 Commit notifications (read-only):
76 Address: multi-tty-commits@lists.fnord.hu
77 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty-commits
83 The branch is now very stable and almost full-featured. All of the
84 major problems have been fixed, only a few minor issues remain. (It
85 still needs to be ported to Windows/Mac/DOS, though.) Both multiple
86 tty device support and simultaneous X and tty frame support works
87 fine. Emacsclient has been extended to support opening new tty and X
88 frames. It has been changed open new Emacs frames by default.
90 The multi-tty branch has been scheduled for inclusion in the next
91 major release of Emacs (version 23). I expect the merge into the
92 development trunk to occur sometime during next year (2005), after the
93 merge of the Unicode branch.
95 Tested on GNU/Linux, Solaris 8, FreeBSD and OpenBSD. Please let me
96 know if you succeed or fail to use it on other platforms---I'll have a
97 few tricky test cases for you.
101 * GTK support. If you compile your Emacs with the GTK
102 toolkit, some functionality of multi-tty will be lost. In
103 particular, you will not be able to work on multiple X
104 displays at once. Current releases of GTK have limitations
105 and bugs that prevent full-blown multi-display support in
106 Emacs. (GTK crashes when Emacs tries to disconnect from an
107 X server.) Use the Lucid toolkit if you want to see a
108 complete feature set.
110 * The single-kboard mode.
112 If your multi-tty Emacs session seems to be frozen, you
113 probably have a recursive editing session or a pending
114 minibuffer prompt (which is a kind of recursive editing) on
115 another display. To unfreeze your session, switch to that
116 display and complete the recursive edit, for example by
117 pressing C-] (`abort-recursive-edit').
119 I am sorry to say that currently there is no way to break
120 out of this "single-kboard mode" from a frozen display. If
121 you are unable to switch to the display that locks the
122 others (for example because it is on a remote computer),
123 then you can use emacsclient to break out of all recursive
126 emacsclient -e '(top-level)'
128 Note that this (perhaps) unintuitive behaviour is by design.
129 Single-kboard mode is required because of an intrinsic Emacs
130 limitation that is very hard to eliminate. (This limitation
131 is related to the single-threaded nature of Emacs.)
133 I plan to implement better user notification and support for
134 breaking out of single-kboard mode from locked displays.
136 * Mac, Windows and DOS support is broken, doesn't even
137 compile. Multiple display support will probably not provide
138 new Emacs features on these systems, but the multi-tty
139 branch changed a few low-level interfaces, and the
140 system-dependent source files need to be adapted
141 accordingly. The changes are mostly trivial, so almost
142 anyone can help, if only by compiling the branch and
143 reporting the compiler errors. (It is not worth to do this
147 HOW TO GET THE BRANCH
148 ---------------------
150 The branch uses GNU Arch (http://www.gnuarch.org) for version control.
152 Retrieving the latest version of the branch:
154 tla register-archive -f http://lorentey.hu/arch/2004/
155 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
157 This incantation uses my private archive mirror that is hosted on a
158 relatively low-bandwidth site; if you are outside Hungary, you will
159 probably want to you use my secondary mirror: (Note that the -f option
160 will overwrite the archive location if you have previously registered
163 tla register-archive -f http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
164 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
166 http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
168 The Arch supermirror provides mirroring services for all public Arch
169 repositories. We have a mirror there, too, if you prefer.
171 tla register-archive -f http://mirrors.sourcecontrol.net/lorentey%40elte.hu--2004
172 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
174 My GPG key id is 0FB27A3F; it is available from
175 hkp://wwwkeys.eu.pgp.net/, or from my homepage at
176 http://lorentey.hu/rolam/gpg.html)
178 Don't worry if the above checkout takes a few minutes to complete;
179 once you have a source tree, updating it to the latest revision will
180 be _much_ faster. Use the following command for the update:
184 You can find more information about Arch on http://wiki.gnuarch.org/.
185 It's a wonderful source control system, I highly recommend it.
187 If you don't have tla, the branch has a homepage from which you can
188 download conventional patches against Emacs CVS HEAD:
190 http://lorentey.hu/project/emacs.html
195 If you run Debian, or a distribution based on Debian, you are welcome
196 to use our binary packages; put these lines in your /etc/apt/sources.list:
199 deb http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
200 deb-src http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
202 Note that these packages are intended solely to provide an easy way to
203 test the new multi-tty features. They are not to be taken as Emacs
204 releases, and it's a mistake to expect robust operation or any kind of
205 timely support for them. Do not install them, unless you'd like to
206 have your editor crash on you.
212 The multi-tty branch is compiled the same way as Emacs itself:
214 make maintainer-clean # (If you have compiled Emacs before)
216 ./configure --without-gtk <your favourite options>
220 If you have strange compilation errors, they may be caused by old
221 *.elc files that are left over from an earlier bootstrap. The `make
222 maintainer-clean' target deletes them, so it is a good idea to run
223 that before reporting a bug. (Emacs requires a clean recompilation
224 after certain kinds of source code changes.)
229 To test the multi-tty branch, start up the Emacs server with the
235 and then (from a shell prompt on another terminal) start emacsclient
237 emacsclient -t /optional/file/names... (for a tty frame)
238 emacsclient /optional/file/names... (for an X frame)
240 (Make sure both emacs and emacsclient are multi-tty versions.)
241 You'll hopefully have two fully working, independent frames on
242 separate terminals. The new frame is closed automatically when you
243 finish editing the specified files (C-x #), but delete-frame (C-x 5 0)
244 also works. Of course, you can create frames on more than two tty
247 Creating new frames on the same tty with C-x 5 2 works, and they
248 behave the same way as in previous Emacs versions. If you exit emacs,
249 all terminals should be restored to their previous states.
251 This is work in progress, and probably full of bugs. It is a good
252 idea to run emacs from gdb, so that you'll have a live instance to
253 debug if something goes wrong. Please send me your bug reports on our
254 mailing list: multi-tty@lists.fnord.hu
259 I think the best way to use the new Emacs is to have it running inside
260 a disconnected GNU screen session, and always use emacsclient for
261 normal work. One advantage of this is that not a single keystroke of
262 your work will be lost if the display device that you are using
263 crashes, or the network connection times out, or whatever. (I had an
264 extremely unstable X server for some time while I was developing these
265 patches, and running Emacs this way has saved me a number of M-x
266 recover-session invocations.)
268 I use the following two bash scripts to handle my Emacs sessions:
270 -------------------------------------------------------connect-emacs--
272 # Usage: connect-emacs <name> <args>...
274 # Connects to the Emacs instance called NAME. Starts up the instance
275 # if it is not already running. The rest of the arguments are passed
281 if [ -z "$name" ]; then
282 echo "Usage: connect_emacs <name> <args>..." >&2
285 preload-emacs "$name" wait
286 /usr/bin/emacsclient.emacs-multi-tty -s "$name" "$@"
287 ----------------------------------------------------------------------
289 -------------------------------------------------------preload-emacs--
291 # Usage: preload-emacs <name> [<waitp>]
293 # Preloads the Emacs instance called NAME in a detached screen
294 # session. Does nothing if the instance is already running. If WAITP
295 # is non-empty, the function waits until the server starts up and
296 # creates its socket; otherwise it returns immediately.
300 screendir="/var/run/screen/S-$USER"
301 serverdir="/tmp/emacs$UID"
302 emacs=/usr/bin/emacs-multi-tty # Or wherever you installed your multi-tty Emacs
304 if [ -z "$name" ]; then
305 echo "Usage: preload_emacs <name> [<waitp>]" >&2
309 if [ ! -e "$screendir"/*."$name" ]; then
310 if [ -e "$serverdir/$name" ]; then
311 # Delete leftover socket (for the wait option)
312 rm "$serverdir/$name"
314 screen -dmS "$name" "$emacs" -nw --eval "(setq server-name \"$name\")" -f server-start
316 if [ ! -z "$waitp" ]; then
317 while [ ! -e "$serverdir/$name" ]; do sleep 0.1; done
319 ----------------------------------------------------------------------
321 I have the following in my profile to have two instances automatically
322 preloaded for editing and email:
327 It is useful to set up short aliases for connect-emacs. I use the
330 alias edit="connect-emacs editor"
332 alias et="connect-emacs editor -t"
333 alias gnus="connect-emacs gnus"
345 For the NEWS file: (Needs much, much work)
347 ** Support for multiple terminal devices has been added.
349 *** You can specify a terminal device (`tty' parameter) and a terminal
350 type (`tty-type' parameter) to `make-terminal-frame'.
352 *** You can test for the presence of multiple terminal support by
353 testing for the `multi-tty' feature.
355 *** Emacsclient has been extended to support opening a new terminal
356 frame. Its behaviour has been changed to open a new Emacs frame by
357 default. Use the -c option to get the old behavior of opening
358 files in the currently selected Emacs frame.
360 *** A make-frame-on-tty function has been added to make it easier to
361 create frames on new terminals.
363 *** New functions: frame-tty-name, frame-tty-type, delete-tty,
364 terminal-local-value, set-terminal-local-value
366 terminal-id, terminal-parameters, terminal-parameter,
367 set-terminal-parameter
369 *** New variables: global-key-translation-map
371 *** The keymaps key-translation-map and function-key-map are now
374 ** Support for simultaneous graphical and terminal frames has been
377 *** The function `make-frame-on-display' now works during a terminal
378 session, and `make-frame-on-tty' works during a graphical session.
380 *** The `window-system' variable has been made frame-local.
382 *** The new `initial-window-system' variable contains the
383 `window-system' value for the first frame.
385 *** talk.el has been extended for multiple tty support.
390 (The rest of this file consists of my development notes and as such it
391 is probably not very interesting for anyone else.)
396 ** Emacs crashes when a tty frame is resized so that there is no space
397 for all its windows. (Tom Schutzer-Weissmann)
399 ** Report GTK multi-display problems to GTK maintainers. For extra
402 Currently you can not connect to new X displays when you compile
403 Emacs with GTK support. If you want to play around with GTK
404 multidisplay (and don't mind core dumps), you can edit src/config.h
405 and define HAVE_GTK_MULTIDISPLAY there by hand.
407 ** Audit `face-valid-attribute-values' usage in customize and
408 elsewhere. Its return value depends on the current window system.
409 Replace static initializers using it with runtime functions. For
410 example, custom's buttons are broken on non-initial device types.
412 ** Possibly turn off the double C-g feature when there is an X frame.
413 C.f. (emacs)Emergency Escape.
415 ** frames-on-display-list should also accept frames.
417 ** I smell something funny around pop_kboard's "deleted kboard" case.
418 Determine what are the circumstances of this case, and fix any
419 bug that comes to light.
421 ** Consider the `tty-type' frame parameter and the `display-tty-type'
422 function. They serve the exact same purpose. I think it may be
423 a good idea to eliminate one of them, preferably `tty-type'.
425 ** The handling of lisp/term/*.el, and frame creation in general, is a
426 big, big mess. How come the terminal-specific file is loaded by
427 tty-create-frame-with-faces? I don't think it is necessary to load
428 these files for each frame; once per terminal should be enough.
429 Update: lisp/term/*.el is not loaded repeatedly anymore, but
430 faces.el still needs to be cleaned up.
432 ** Fix frame-set-background-mode in this branch. It was recently
433 changed in CVS, and frame.el in multi-tty has not yet been adapted
434 for the changes. (It needs to look at
435 default-frame-background-mode.) (Update: maybe it is fixed now;
438 ** I think `(set-)terminal-local-value' and the terminal parameter
439 mechanism should be integrated into a single framework.
441 ** Add the following hooks: after-delete-frame-hook (for server.el,
442 instead of delete-frame-functions),
443 after-delete-terminal-functions, after-create-terminal-functions.
445 ** Fix set-input-mode for multi-tty. It's a truly horrible interface;
446 what if we'd blow it up into several separate functions (with a
447 compatibility definition)?
449 ** BULK RENAME: The `display-' prefix of new Lisp-level functions
450 conflicts with stuff like `display-time-mode'. Use `device-'
451 or `terminal-' instead. I think I prefer `terminal-'.
453 It turns out that most of the offending Lisp functions were defined
454 in the trunk. Therefore, compatibility aliases should be defined
455 for the following names:
457 display-color-cells terminal-color-cells
458 display-color-p terminal-color-p
459 display-graphic-p terminal-graphic-p
460 display-grayscale-p terminal-grayscale-p
461 display-images-p terminal-images-p
462 display-mm-height terminal-mm-height
463 display-mm-width terminal-mm-width
464 display-mouse-p terminal-mouse-p
465 display-multi-font-p terminal-multi-font-p
466 display-multi-frame-p terminal-multi-frame-p
467 display-pixel-height terminal-pixel-height
468 display-pixel-width terminal-pixel-width
469 display-pixels-per-inch terminal-pixels-per-inch
470 display-planes terminal-planes
471 display-popup-menus-p terminal-popup-menus-p
472 display-save-under terminal-save-under
473 display-screens terminal-screens
474 display-supports-face-attributes-p terminal-supports-face-attributes-p
475 display-visual-class terminal-visual-class
476 framep-on-display framep-on-terminal
477 frames-on-display-list frames-on-terminal-list
479 The following functions were introduced in the multi-tty branch, and
480 can be renamed without aliases:
482 display-controlling-tty-p terminal-controlling-tty-p
483 display-list terminal-list
484 display-live-p terminal-live-p
485 display-name terminal-name
486 display-tty-type terminal-tty-type
487 frame-display terminal-of-frame
488 delete-display delete-terminal
490 ** The single-keyboard mode of MULTI_KBOARD is extremely confusing
491 sometimes; Emacs does not respond to stimuli from other keyboards.
492 At least a beep or a message would be important, if the single-mode
493 is still required to prevent interference. (Reported by Dan
496 Update: selecting a region with the mouse enables single_kboard
497 under X. This is very confusing.
499 Update: After discussions with Richard Stallman, this will be
500 resolved by having locked displays warn the user to wait, and
501 introducing a complex protocol to remotely bail out of
502 single-kboard mode by pressing C-g.
504 Update: Warning the user is not trivial to implement, as Emacs has
505 only one echo area, shared by all frames. Ideally the warning
506 should not be displayed on the display that is locking the others.
507 Perhaps the high probability of user confusion caused by
508 single_kboard mode deserves a special case in the display code.
509 Alternatively, it might be good enough to signal single_kboard mode
510 by changing the modelines or some other frame-local display element
511 on the locked out displays.
513 ** The session management module is prone to crashes when the X
514 connection is closed and then later I try to connect to a new X
517 #0 0xb7ebc806 in SmcGetIceConnection () from /usr/X11R6/lib/libSM.so.6
518 #1 0x080e6641 in x_session_check_input (bufp=0xbf86c9c0) at xsmfns.c:144
519 #2 0x080d3bbc in XTread_socket (device=0xa722ff8, expected=1, hold_quit=0xbf86ca90) at xterm.c:7037
520 #3 0x080fa404 in read_avail_input (expected=1) at keyboard.c:6696
521 #4 0x080fa4ca in handle_async_input () at keyboard.c:6900
522 #5 0x080d51fa in x_term_init (display_name=162628899, xrm_option=0x0, resource_name=0x857068c "emacs") at xterm.c:10622
523 #6 0x080d920e in x_display_info_for_name (name=162628899) at xfns.c:3975
524 #7 0x080d92f9 in check_x_display_info (object=1) at xfns.c:274
525 #8 0x080d97b8 in Fx_create_frame (parms=151221485) at xfns.c:3016
526 #9 0x0815bf72 in Ffuncall (nargs=2, args=0xbf86ceec) at eval.c:2851
528 I installed a workaround to prevent this. The X session manager is
529 only contacted when the very first display in the Emacs session is
530 an X display. Also, x_delete_display() on this display aborts
531 session management, and XTread_socket only calls
532 x_session_check_input when it is called for the display that the
533 session was opened on. While this does not really fix the bug, it
534 makes it much less frequent, because session manager support will
535 not normally be enabled when Emacs can survive the shutdown of the
538 See if xsmfns.c should be updated.
540 ** Hunt down display-related functions in frame.el and extend them all
541 to accept display ids.
543 ** rif->flush_display_optional (NULL) calls should be replaced by a
546 ** The set-locale-environment hack (adding the DISPLAY option) should
547 be replaced with a clean design.
549 ** standard-display-table should be display-local.
550 standard-display-european should be display-local.
552 ** With iswitchb-default-method set to 'always-frame, only frames on
553 the current display should be considered. This might involve
554 extending `get-buffer-window'.
556 ** Have a look at Vlocale_coding_system. Seems like it would be a
557 tedious job to localize it, although most references use it for
558 interfacing with libc and are therefore OK with the global
561 Exceptions found so far: x-select-text and
562 x-cut-buffer-or-selection-value.
564 ** Have a look at fatal_error_hook.
566 ** Have a look at set_frame_matrix_frame.
568 ** Check if we got term-setup-hook right.
570 ** I think tip_frame should be display-local.
572 ** Check display reference count handling in x_create_tip_frame.
574 ** make-frame does not correctly handle extra parameters in its
577 (frame-parameter (make-frame (list (cons 'foobar 42))) 'foobar)
580 (This is likely an error in the CVS trunk.)
582 ** Dan Nicolaescu suggests that -nw should be added as an alias for -t
583 in emacsclient. Good idea. (Alas, implementing this is not
584 trivial, getopt_long does not seem to support two-letter ``short''
585 options. Patches are welcome.)
587 ** Mark Plaksin suggests that emacsclient should accept the same
588 X-related command-line arguments as Emacs. Most of the X-related
589 argument-handling is done in Lisp, so this should be quite easy to
592 ** Gergely Nagy suggests that C-x # should only kill the current
593 frame, not any other emacsclient frame that may have the same file
594 opened for editing. I think I agree with him.
596 ** Very strange bug: visible-bell does not work on secondary
597 terminals in xterm and konsole. The screen does flicker a bit,
598 but it's so quick it isn't noticable.
600 ** Move baud_rate to struct display.
602 ** Implement support for starting an interactive Emacs session without
603 an initial frame. (The user would connect to it and open frames
604 later, with emacsclient.)
606 ** Fix Mac support (I can't do this entirely myself). Note that the
607 current state of Mac-specific source files in the multi-tty tree
608 are not useful; before starting work on Mac support, revert to
609 pristine, pre-multi-tty versions.
611 ** Fix W32 support (I can't do this entirely myself). Note that the
612 current state of W32-specific source files in the multi-tty tree
613 are not useful; before starting work on W32 support, revert to
614 pristine, pre-multi-tty versions.
616 ** Fix DOS support (I can't do this entirely myself). Note that the
617 current state of DOS-specific source files in the multi-tty tree
618 are not useful; before starting work on DOS support, revert to
619 pristine, pre-multi-tty versions.
621 ** Do a grep on XXX and ?? for more issues.
623 ** Understand Emacs's low-level input system (it's black magic) :-)
624 What exactly does interrupt_input do? I tried to disable it for
625 raw secondary tty support, but it does not seem to do anything
626 useful. (Update: Look again. X unconditionally enables this, maybe
627 that's why raw terminal support is broken again. I really do need
628 to understand input.)
629 (Update: I am starting to understand the read_key_sequence->read-char
630 ->kbd_buffer_get_event->read_avail_input->read_socket_hook path. Yay!)
632 ** flow-ctrl.el must be updated.
634 ** Fix stuff_char for multi-tty. Doesn't seem to be of high priority.
639 (ex-TODO items with explanations.)
641 -- Introduce a new struct for terminal devices.
643 (Done, see struct tty_output. The list of members is not yet
646 -- Change the bootstrap procedure to initialize tty_list.
648 (Done, but needs review.)
650 -- Change make-terminal-frame to support specifying another tty.
652 (Done, new frame parameters: `tty' and `tty-type'.)
654 -- Implement support for reading from multiple terminals.
656 (Done, read_avail_input tries to read from each terminal, until one
657 succeeds. MULTI_KBOARD is not used. Secondary terminals don't send
660 (Update: They do, now.)
662 (Update2: After enabling X, they don't.)
664 -- other-frame should cycle through the frames on the `current'
667 (Done, by trivially modifiying next_frame and prev_frame.)
669 -- Support different terminal sizes.
673 -- Make sure terminal resizes are handled gracefully. (Could be
676 (Done. We don't get automatic SIGWINCH for additional ttys,
679 -- Extend emacsclient to automatically open a new tty when it connects
682 (Done. It's an ugly hack, needs more work.)
684 -- Redisplay must refresh the topmost frame on *all* terminals, not
685 just the initial terminal.
687 (Done, but introduced an ugly redisplay problems. Ugh.)
689 -- Fix redisplay problems.
691 (Done; it turned out that the entire Wcm structure must be moved
692 inside tty_output. Why didn't I catch this earlier?)
694 -- Provide a way for emacsclient to tell Emacs that the tty has been
697 (Done, simply forward the SIGWINCH signal.)
699 -- Each keypress should automatically select the frame corresponding
700 to the terminal that it was coming from. This means that Emacs
701 must know from which terminal the last keyboard event came from.
703 (Done, it was quite simple, the input event system already
704 supported multiple frames.)
706 -- Fix SIGIO issue with secondary terminals.
708 (Done, emacsclient signals Emacs after writing to the proxy pseudo
709 terminal. Note that this means that multi-tty does not work with
712 (Update: This is bullshit. There is a read_input_waiting function,
713 extend that somehow.)
715 (Update of update: The first update was not right either, extending
716 read_input_waiting was not necessary. Secondary ttys do seem to
717 send signals on input.)
719 (Update^3: Not any more.)
721 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
722 parameters from the currently selected terminal before the global
727 -- Put all cached terminal escape sequences into struct tty_output.
728 Currently, they are still stored in global variables, so we don't
729 really support multiple terminal types.
731 (Done. It was not fun.)
733 -- Implement sane error handling after initialization. (Currently
734 emacs exits if you specify a bad terminal type.) The helpful error
735 messages must still be provided when Emacs starts.
739 -- Implement terminal deletion, i.e., deleting local frames, closing
740 the tty device and restoring its previous state without exiting
743 (Done, but at the moment only called when an error happens during
744 initialization. There is a memory corruption error around this
745 somewhere.) (Update: now it is fully enabled.)
747 -- Implement automatic deletion of terminals when the last frame on
748 that terminal is closed.
752 -- Restore tty screen after closing the terminal.
754 (Done, we do the same as Emacs 21.2 for all terminals.)
756 -- 'TERM=dumb src/emacs' does not restore the terminal state.
760 -- C-g should work on secondary terminals.
762 (Done, but the binding is not configurable.)
764 -- Deal with SIGHUP in Emacs and in emacsclient. (After this, the
765 server-frames may be removed from server.el.)
767 (Done, nothing to do. It seems that Emacs does not receive SIGHUP
768 from secondary ttys, which is actually a good thing.) (Update: I
769 think it would be a bad idea to remove server-frames.)
771 -- Change emacsclient/server.el to support the -t argument better,
772 i.e. automatically close the socket when the frame is closed.
774 (Seems to be working OK.)
776 -- Fix mysterious memory corruption error with tty deletion. To
777 trigger it, try the following shell command:
779 while true; do TERM=no-such-terminal-definition emacsclient -h; done
781 Emacs usually dumps core after a few dozen iterations. (The bug
782 seems to be related to the xfreeing or bzeroing of
783 tty_output.Wcm. Maybe there are outside references to struct Wcm?
784 Why were these vars collected into a struct before multi-tty
787 (Done. Whew. It turned out that the problem had nothing to do
788 with hypothetical external references to Wcm, or any other
789 tty_output component; it was simply that delete_tty closed the
790 filehandles of secondary ttys twice, resulting in fclose doubly
791 freeing memory. Utterly trivial matter. I love the C's memory
792 management, it puts hair on your chest.)
794 -- Support raw secondary terminals. (Note that SIGIO works only on
795 the controlling terminal.) Hint: extend read_input_waiting for
796 multiple ttys and hopefully this will be fixed.
798 (Done, it seems to have been working already for some time. It
799 seems F_SETOWN does work, after all. Not sure what made it fail
800 earlier, but it seems to be fixed (there were several changes
801 around request_sigio, maybe one of them did it).
802 read_input_waiting is only used in sys_select, don't change
803 it.) (Update: After adding X support, it's broken again.)
804 (Update^2: No it isn't.) :-)
806 -- Find out why does Emacs abort when it wants to close its
807 controlling tty. Hint: chan_process[] array. Hey, maybe
808 noninterrupt-IO would work, too? Update: no, there is no process
811 (Done. Added add/delete_keyboard_wait_descriptor to
812 term_init/delete_tty. The hint was right, in a way.)
814 -- Issue with SIGIO: it needs to be disabled during redisplay. See if
815 fcntl kernel behaviour could be emulated by emacsclient.
817 (Done. Simply disabled the SIGIO emulation hack in emacsclient.)
818 (Update: it was added back.) (Update^2: and removed again.)
820 -- server.el: There are issues with saving files in buffers of closed
821 clients. Try editing a file with emacsclient -f, and (without
822 saving it) do a delete-frame. The frame is closed without
823 question, and a surprising confirmation prompt appears in another
826 (Done. delete-frame now asks for confirmation if it still has
827 pending buffers, and modified buffers don't seem to be deleted.)
829 -- emacsclient.el, server.el: Handle eval or file open errors when
834 -- Make parts of struct tty_output accessible from Lisp. The device
835 name and the type is sufficient.
837 (Done, see frame-tty-name and frame-tty-type.)
839 -- Export delete_tty to the Lisp environment, for emacsclient.
841 (Done, see delete-tty.)
843 -- Get rid of the accessor macros in termchar.h, or define macros for
848 -- Move device-specific parameters (like costs) commonly used by
849 device backends to a common, device-dependent structure.
851 (Done. See struct display_method in termhooks.h.)
855 (Done. Well, it seems to be working.)
857 -- Allow simultaneous X and tty frames. (Handling input could be
858 tricky. Or maybe not.)
860 (Done. Allowed, that is. It is currently extremely unstable, to
861 the point of being unusable. The rif variable causes constant
862 core dumps. Handling input is indeed tricky.)
864 -- Rewrite multi-tty input in terms of MULTI_KBOARD.
866 (Done. In fact, there was no need to rewrite anything, I just
867 added a kboard member to tty_display_info, and initialized the
868 frame's kboard from there.)
870 -- Fix rif issue with X-tty combo sessions. IMHO the best thing to do
871 is to get rid of that global variable (and use the value value in
872 display_method, which is guaranteed to be correct).
874 (Done, did exactly that. Core dumps during combo sessions became
875 much rarer. In fact, I have not yet met a single one.)
877 -- Add multi-tty support to talk.el.
881 -- Clean up the source of emacsclient. It is a mess.
883 (Done, eliminated stupid proxy-pty kludge.)
885 -- Fix faces on tty frames during X-tty combo sessions. There is an
886 init_frame_faces call in init_sys_modes, see if there is a problem
889 (Done, there was a stupid mistake in
890 Ftty_supports_face_attributes_p. Colors are broken, though.)
892 -- C-x 5 2, C-x 5 o, C-x 5 0 on an emacsclient frame unexpectedly
893 exits emacsclient. This is a result of trying to be clever with
894 delete-frame-functions.
896 (Fixed, added delete-tty-after-functions, and changed server.el to
899 -- Something with (maybe) multi-keyboard support broke function keys
900 and arrows on ttys during X+tty combo sessions. Debug this.
902 (I can't reproduce it, maybe the terminal type was wrong.)
904 -- Fix input from raw ttys (again).
906 (Now it seems to work all right.)
908 -- During an X-tty combo session, a (message "Hello") from a tty frame
909 goes to the X frame. Fix this.
911 (Done. There was a safeguard against writing to the initial
912 terminal frame during bootstrap which prevented echo_area_display
913 from working correctly on a tty frame during a combo session.)
915 -- If there are no frames on its controlling terminal, Emacs should
916 exit if the user presses C-c there.
918 (Done, as far as possible. See the SIGTERM comment in
919 interrupt_signal on why this seems to be impossible to solve this
922 -- During an X session, Emacs seems to read from stdin. Also, Emacs
923 fails to start without a controlling tty.
925 (Fixed by replacing the troublesome termcap display with a dummy
926 bootstrap display during bootstrap.
928 -- Do tty output through struct display, like graphical display
933 -- Define an output_initial value for output_method for the initial
934 frame that is dumped with Emacs. Checking for this frame (e.g. in
935 cmd_error_internal) is ugly.
937 (Done, breaking interactive temacs.)
939 -- The command `emacsclient -t -e '(delete-frame)'' fails to exit.
943 -- frame-creation-function should always create a frame that is on the
944 same display as the selected frame. Maybe frame-creation-function
945 should simply be removed and make-frame changed to do the right
948 (Done, with a nice hack. frame-creation-function is now frame-local.)
950 -- Fix C-g on raw ttys.
952 (Done. I disabled the interrupt/quit keys on all secondary
953 terminals, so Emacs sees C-g as normal input. This looks like an
954 overkill, because emacsclient has extra code to pass SIGINT to
955 Emacs, so C-g should remain the interrupt/quit key on emacsclient
956 frames. See the next entry why implementing this distinction would
959 -- Make sure C-g goes to the right frame with ttys. This is hard, as
960 SIGINT doesn't have a tty parameter. :-(
962 (Done, the previous change fixes this as a pleasant side effect.)
964 -- I have seen a case when Emacs with multiple ttys fell into a loop
965 eating 100% of CPU time. Strace showed this loop:
968 kill(30284, SIGIO) = 0
969 --- SIGIO (I/O possible) @ 0 (0) ---
970 ioctl(6, FIONREAD, [0]) = -1 EIO (Input/output error)
971 ioctl(5, FIONREAD, [0]) = -1 EIO (Input/output error)
972 ioctl(0, FIONREAD, [0]) = 0
973 sigreturn() = ? (mask now [])
974 gettimeofday({1072842297, 747760}, NULL) = 0
975 gettimeofday({1072842297, 747806}, NULL) = 0
976 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
977 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
978 gettimeofday({1072842297, 748245}, NULL) = 0
980 I have seen something similar with a single X frame, but have not
981 been able to reproduce it for debugging.
983 Update: This may have been caused by checking for nread != 0
984 instead of nread > 0 after calling read_socket_hook in
987 (Fixed. This was caused by unconditionally including stdin in
988 input_wait_mask in init_process. The select call in
989 wait_reading_process_input always returned immediately, indicating
990 that there is pending input from stdin, which nobody read.
992 Note that the above strace output seems to be an unrelated but
993 similar bug. I think that is now fixed.)
995 -- Exiting Emacs while there are emacsclient frames doesn't restore the
996 ttys to their default states.
998 (This seems to be fixed by some previous change.)
1000 -- Allow opening an X session after -nw.
1004 -- Fix color handling during tty+X combo sessions. (It seems that tty
1005 sessions automatically convert the face colors to terminal colors
1006 when the face is loaded. This conversion must happen instead on
1007 the fly in write_glyphs, which might be problematic, as color
1008 approximation is currently done in lisp (term/tty-colors.el).)
1009 (Update: hm, colors seem to work fine if I start emacs with -nw and
1010 then create an X frame. Maybe it's just a small buglet somewhere.)
1012 (Seems to be fixed. The problem was in startup.el, it did not
1013 initialize tty colors when the initial window system was
1016 -- emacs -nw --eval '(y-or-n-p "Foobar")' segfaults. (Reported by
1019 (Fixed, there was a keyboard initialization problem.)
1021 -- Fix interactive use of temacs. There are face-related SEGVs, most
1022 likely because of changes in realize_default_face, realize_face.
1026 -- Don't exit Emacs when the last X connection fails during a
1027 multi-display session.
1031 -- Dan Nicolaescu noticed that starting emacsclient on the same
1032 terminal device that is the controlling tty of the Emacs process
1033 gives unexpected results.
1037 -- Istvan Marko reported that Emacs hang on ttys if it was started
1038 from a shell script.
1040 (Fixed. There was a bug in the multi-tty version of
1041 narrow_foreground_group. tcsetpgrp blocks if it is called from a
1042 process that is not in the same process group as the tty.)
1044 -- emacsclient -t from an Emacs term buffer does not work, complains
1045 about face problems. This can even lock up Emacs (if the recursive
1046 frame sets single_kboard). Update: the face problems are caused by
1047 bugs in term.el, not in multi-tty. The lockup is caused by
1048 single_kboard mode, and is not easily resolvable. The best thing to
1049 do is to simply refuse to create a tty frame of type `eterm'.
1051 (Fixed, changed emacsclient to check for TERM=eterm. The face
1052 complaints seem to be caused by bugs in term.el; they are not
1053 related to multi-tty.)
1055 -- Find out the best way to support suspending Emacs with multiple
1056 ttys. My guess: disable it on the controlling tty, but from other
1057 ttys pass it on to emacsclient somehow. (It is (I hope) trivial to
1058 extend emacsclient to handle suspend/resume. A `kill -STOP' almost
1061 (Done. I needed to play with signal handling and the server
1062 protocol a bit to make emacsclient behave as a normal UNIX program
1063 wrt foreground/background process groups.)
1065 -- There is a flicker during the startup of `emacs -nw'; it's as if
1066 the terminal is initialized, reset and then initialialized again.
1067 Debug this. (Hint: narrow_foreground_group is called twice during
1072 -- Robert Chassell has found serious copy-paste bugs with the
1073 multi-tty branch. There seem to be redisplay bugs while copying
1074 from X to a terminal frame. Copying accented characters do not
1077 (Patch-124 should fix this, by changing the interprogram-*-function
1078 variables to be frame-local, as suggested by Mark Plaksin
1079 (thanks!). I think that the redisplay bugs are in fact not bugs,
1080 but delays caused by single_kboard --> perhaps MULTI_KBOARD should
1083 -- frame-creation-function was removed, which might be a bad idea.
1084 Think up a compatible solution.
1086 (It was an internal interface that may be changed when necessary.)
1088 -- Change Lisp code not to (getenv "TERM"); use the `tty-type' frame
1089 parameter or the frame-tty-type function instead. (M-x tags-search
1090 "TERM" helps with this.) Update: Actually, all getenv invocations
1091 should be checked for multi-tty compatibility, and an interface
1092 must be implemented to get the remote client's environment.
1094 (Done. Only getenv calls in lisp/term/*.el were changed; other
1095 calls should be mostly left as they are.)
1097 -- Add an elaborate mechanism for display-local variables. (There are
1098 already a few of these; search for `terminal-local' in the Elisp
1101 (Not needed. Display-local variables could be emulated by
1102 frame-local variables.)
1104 -- Emacs assumes that all terminal frames have the same locale
1105 settings as Emacs itself. This may lead to bogus results in a
1106 multi-locale setup. (E.g., while logging in from a remote client
1107 with a different locale.)
1108 (Update after new bugreport by Friedrich Delgado Friedrichs:
1109 (at least) the structs terminal_coding and keyboard_coding in
1110 coding.c must be moved to struct display, and the Lisp interface
1111 [set-]keyboard-coding-system must be adapted for the change.)
1113 (Fixed. Emacs now uses the locale settings as seen by the
1114 emacsclient process for server tty frames.)
1115 (Update: Not really; Vlocale_coding_system is still global.)
1117 -- Make `struct display' accessible to Lisp programs. Accessor functions:
1119 (displayp OBJECT): Returns t if OBJECT is a display.
1120 => Implemented as display-live-p.
1122 (display-list): Returns list of currently active displays.
1125 (selected-display): Returns the display object of the selected frame.
1126 => Not strictly necessary, but implemented anyway.
1128 (frame-display FRAME): Returns the display object of FRAME.
1131 (display-frames DISPLAY): Returns a list of frames on DISPLAY.
1132 => Already implemented, see frames-on-display-list.
1134 (display-type DISPLAY): Returns the type of DISPLAY, as a
1135 symbol. (See `framep'.)
1136 => Implemented as display-live-p.
1138 (display-device DISPLAY): Returns the name of the device that
1139 DISPLAY uses, as a string. (E.g: "/dev/pts/16", or
1141 => Implemented as display-name.
1145 See next issue why this is necessary.
1147 (Update: The consensus on emacs-devel seems to be to do this via
1148 integer identifiers. That's fine by me.)
1152 -- The following needs to be supported:
1159 The cleanest way to solve this is to allow multiple displays on the
1160 same terminal device; each new emacsclient process should create
1161 its own display. As displays are currently identified by their
1162 device names, this is not possible until struct display becomes
1163 accessible as a Lisp-level object.
1167 -- Miles Bader suggests that C-x C-c on an emacsclient frame should
1168 only close the frame, not exit the entire Emacs session. Update:
1169 see above for a function that does this. Maybe this should be the
1172 (Done. This is the new default. No complaints so far.)
1174 -- Clean up the frame-local variable system. I think it's ugly and
1175 error-prone. But maybe I just haven't yet fully understood it.
1177 (Nothing to do. It doesn't seem ugly any more. It's rather clever.)
1179 -- Support multiple character locales. A version of
1180 `set-locale-environment' needs to be written for setting up
1181 display-local settings on ttys. I think calling
1182 set-display-table-and-terminal-coding-system and
1183 set-keyboard-coding-system would be enough. The language
1184 environment itself should remain a global setting.
1186 (Done, by an ugly hack.)
1188 -- The terminal customization files in term/*.el tend to change global
1189 parameters, which may confuse Emacs with multiple displays. Change
1190 them to tweak only frame-local settings, if possible. (They tend
1191 to call define-key to set function key sequences a lot.)
1193 (Done, by making `function-key-map' terminal-local (i.e., part of
1194 struct kboard). This has probably covered all the remaining problems.)
1196 -- Make `function-key-map' and `key-translation-map' terminal-local.
1200 -- Implement `terminal-local-value' and `set-terminal-local-value' to
1201 allow deterministic access to terminal local bindings. The
1202 encode-kb package can not set up `key-translation-map' without
1203 these. The terminal-local bindings seem to be independent of what
1208 -- xt-mouse.el needs to be adapted for multi-tty. It currently
1209 signals an error on kill-emacs under X, which prevents the user
1210 from exiting Emacs. (Reported by Mnemonikk on freenode.)
1215 -- Having {reset,init}_all_sys_modes in set-input-mode breaks arrow
1216 keys on non-selected terminals under screen, and sometimes on other
1217 terminal types as well. The other function keys continue to work
1218 fine. Sometimes faces on these screens become garbled.
1220 This only seems to affect displays that are of the same terminfo
1221 type as the selected one. Interestingly, in screen Emacs normally
1222 reports the up arrow key as `M-o A', but after the above SNAFU, it
1223 complains about `M-[ a'. UNIX ttys are a complete mystery to me,
1224 but it seems the reset-reinitialize cycle somehow leaves the
1225 non-selected terminals in a different state than usual. I have no
1226 idea how this could happen.
1228 Currently set-input-mode resets only the currently selected
1229 terminal, which seems to somehow work around the problem.
1233 Dan Nicolaescu <dann@ics.uci.edu> writes:
1234 > Some terminals have 2 modes for cursor keys: Application Mode where
1235 > the cursor keys transmit the codes defined in the terminfo entry, and
1236 > Cursor mode. Applications have to send the smkx and rmkx terminfo
1237 > strings to switch between the 2 modes. So Emacs (and emacsclient) have
1238 > to send smkx when initializing and rmkx when quitting (or on
1241 (I think patch-370 fixed this.)
1243 -- This long-standing bug (first reported by Han Boetes) seems to come
1244 and go all the time. It is time to track it down and fix it.
1249 # From another xterm:
1250 emacsclient -e '(y-or-n-p "Do you want me to crash? ")'
1251 # Notice how the answer ends up in the *scratch* buffer
1255 (Fixed in patch-414 after detailed analysis by Kalle Olavi Niemitalo.)
1257 -- normal-erase-is-backspace-mode in simple.el needs to be updated for
1258 multi-tty (rep. by Dan Waber). (The Delete key is broken on X
1261 (Fixed in patch-427.)
1263 -- I think keyboard-translate-table should be made terminal-local.
1265 (Done in patch-431.)
1267 -- The semantics of terminal-local variables are confusing; it is not
1268 clear what binding is in effect in any given time. See if
1269 current_kboard (or at least the terminal-local bindings exported to
1270 Lisp) might be changed to be tied to the selected frame instead.
1271 Currently, `function-key-map' and `key-translation-map' may be
1272 accessed reliably only using the hackish
1273 `(set-)terminal-local-value' functions.
1275 Perhaps there should be a difference between `last-command' &co.
1276 and these more conventional configuration variables.
1277 (E.g. `symbol-value' would use current_kboard to access
1278 `last-command', but SELECTED_FRAME()->display->kboard to get the
1279 value of `function-key-map'.
1281 (Fixed in patch-434.)
1283 -- If the first key pressed on a new tty terminal is a function key,
1284 it is not recognized correctly. May be related to the bug below.
1286 (Seems to have been fixed as a side effect of patch-434. "The bug
1287 below" was the set-input-mode madness.)
1289 (Update: this bug was fixed for good in patch-449. It was tracked
1290 down to a bug in `read_key_sequence': it failed to reinitialize its
1291 local function-key-map/key-translation-map references when it
1292 switched keyboards. I don't understand why did this bug only
1293 appear on brand new frames, though!)
1295 -- Disable connecting to a new X display when we use the GTK toolkit.
1297 (Disabled in patch-450.)
1299 ;;; arch-tag: 8da1619e-2e79-41a8-9ac9-a0485daad17d