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 ARISAWA Akihiro <ari@mbf.ocn.ne.jp>
35 Vincent Bernat <bernat@luffy.cx>
36 Han Boetes <han@mijncomputer.nl>
37 Robert J. Chassell <bob@rattlesnake.com>
38 Romain Francoise <romain@orebokech.com>
39 Ami Fischman <ami@fischman.org>
40 Friedrich Delgado Friedrichs <friedel@nomaden.org>
41 IRIE Tetsuya <irie@t.email.ne.jp>
42 Yoshiaki Kasahara <kasahara@nc.kyushu-u.ac.jp>
43 Bas Kok <nekkobassu@yahoo.com>
44 Jurej Kubelka <Juraj.Kubelka@email.cz>
45 David Lichteblau <david@lichteblau.com>
46 Xavier Mallard <zedek@gnu-rox.org>
47 Istvan Marko <mi-mtty@kismala.com>
48 Ted Morse <morse@ciholas.com>
49 Gergely Nagy <algernon@debian.org>
50 Dan Nicolaescu <dann@ics.uci.edu>
51 Kalle Olavi Niemitalo <kon@iki.fi>
52 Mark Plaksin <happy@mcplaksin.org>
53 Francisco Borges <borges@let.rug.nl>
54 Frank Ruell <stoerte@dreamwarrior.net>
55 Dan Waber <dwaber@logolalia.com>
58 Richard Stallman was kind enough to review an earlier version of my
65 The multi-tty mailing list (discussion & bug reports):
67 Address: multi-tty@lists.fnord.hu
68 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty
69 Archive: http://news.gmane.org/gmane.emacs.multi-tty/
71 Commit notifications (read-only):
73 Address: multi-tty-commits@lists.fnord.hu
74 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty-commits
80 The branch is now very stable and almost full-featured. All of the
81 major problems have been fixed, only a few minor issues remain. (It
82 still needs to be ported to Windows/Mac/DOS, though.) Both multiple
83 tty device support and simultaneous X and tty frame support works
84 fine. Emacsclient has been extended to support opening new tty and X
85 frames. It has been changed open new Emacs frames by default.
87 The multi-tty branch has been scheduled for inclusion in the next
88 major release of Emacs (version 23). I expect the merge into the
89 development trunk to occur sometime during next year (2005), after the
90 merge of the Unicode branch.
92 Tested on GNU/Linux, Solaris 8, FreeBSD and OpenBSD. Please let me
93 know if you succeed or fail to use it on other platforms---I'll have a
94 few tricky test cases for you.
98 * GTK support. If you compile your Emacs with the GTK
99 toolkit, some functionality of multi-tty will be lost.
100 In particular, closing an X display causes a crash.
101 Current releases of GTK have limitations and bugs that
102 prevent full-blown multi-display support in Emacs. Use the
103 Lucid toolkit if you want to see a complete feature set.
105 * The single-kboard mode.
107 If your multi-tty Emacs session seems to be frozen, you
108 probably have a recursive editing session or a pending
109 minibuffer prompt (which is a kind of recursive editing) on
110 another display. To unfreeze your session, switch to that
111 display and complete the recursive edit, for example by
112 pressing C-] (`abort-recursive-edit').
114 I am sorry to say that currently there is no way to break
115 out of this "single-kboard mode" from a frozen display. If
116 you are unable to switch to the display that locks the
117 others (for example because it is on a remote computer),
118 then you can use emacsclient to break out of all recursive
121 emacsclient -e '(top-level)'
123 Note that this (perhaps) unintuitive behaviour is by design.
124 Single-kboard mode is required because of an intrinsic Emacs
125 limitation that is very hard to eliminate. (This limitation
126 is related to the single-threaded nature of Emacs.)
128 I plan to implement better user notification and support for
129 breaking out of single-kboard mode from locked displays.
131 * Mac, Windows and DOS support is broken, doesn't even
132 compile. Multiple display support will probably not provide
133 new Emacs features on these systems, but the multi-tty
134 branch changed a few low-level interfaces, and the
135 system-dependent source files need to be adapted
136 accordingly. The changes are mostly trivial, so almost
137 anyone can help, if only by compiling the branch and
138 reporting the compiler errors. (It is not worth to do this
142 HOW TO GET THE BRANCH
143 ---------------------
145 The branch uses GNU Arch (http://www.gnuarch.org) for version control.
147 Retrieving the latest version of the branch:
149 tla register-archive -f http://lorentey.hu/arch/2004/
150 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
152 This incantation uses my private archive mirror that is hosted on a
153 relatively low-bandwidth site; if you are outside Hungary, you will
154 probably want to you use my secondary mirror: (Note that the -f option
155 will overwrite the archive location if you have previously registered
158 tla register-archive -f http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
159 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
161 http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
163 The Arch supermirror provides mirroring services for all public Arch
164 repositories. We have a mirror there, too, if you prefer.
166 tla register-archive -f http://mirrors.sourcecontrol.net/lorentey%40elte.hu--2004
167 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
169 My GPG key id is 0FB27A3F; it is available from
170 hkp://wwwkeys.eu.pgp.net/, or from my homepage at
171 http://lorentey.hu/rolam/gpg.html)
173 Don't worry if the above checkout takes a few minutes to complete;
174 once you have a source tree, updating it to the latest revision will
175 be _much_ faster. Use the following command for the update:
179 You can find more information about Arch on http://wiki.gnuarch.org/.
180 It's a wonderful source control system, I highly recommend it.
182 If you don't have tla, the branch has a homepage from which you can
183 download conventional patches against Emacs CVS HEAD:
185 http://lorentey.hu/project/emacs.html
190 If you run Debian, or a distribution based on Debian, you are welcome
191 to use our binary packages; put these lines in your /etc/apt/sources.list:
194 deb http://aszt.inf.elte.hu/~lorentey/mirror/apt ./
195 deb-src http://aszt.inf.elte.hu/~lorentey/mirror/apt ./
197 Note that these packages are intended solely to provide an easy way to
198 test the new multi-tty features. They are not to be taken as Emacs
199 releases, and it's a mistake to expect robust operation or any kind of
200 timely support for them. Do not install them, unless you'd like to
201 have your editor crash on you.
207 The multi-tty branch is compiled the same way as Emacs itself:
209 make maintainer-clean # (If you have compiled Emacs before)
211 ./configure <your favourite options>
215 If you have strange compilation errors, they may be caused by old
216 *.elc files that are left over from an earlier bootstrap. The `make
217 maintainer-clean' target deletes them, so it is a good idea to run
218 that before reporting a bug. (Emacs requires a clean recompilation
219 after certain kinds of source code changes.)
224 To test the multi-tty branch, start up the Emacs server with the
230 and then (from a shell prompt on another terminal) start emacsclient
232 emacsclient -t /optional/file/names... (for a tty frame)
233 emacsclient /optional/file/names... (for an X frame)
235 (Make sure both emacs and emacsclient are multi-tty versions.)
236 You'll hopefully have two fully working, independent frames on
237 separate terminals. The new frame is closed automatically when you
238 finish editing the specified files (C-x #), but delete-frame (C-x 5 0)
239 also works. Of course, you can create frames on more than two tty
242 Creating new frames on the same tty with C-x 5 2 works, and they
243 behave the same way as in previous Emacs versions. If you exit emacs,
244 all terminals should be restored to their previous states.
246 This is work in progress, and probably full of bugs. It is a good
247 idea to run emacs from gdb, so that you'll have a live instance to
248 debug if something goes wrong. Please send me your bug reports on our
249 mailing list: multi-tty@lists.fnord.hu
254 I think the best way to use the new Emacs is to have it running inside
255 a disconnected GNU screen session, and always use emacsclient for
256 normal work. One advantage of this is that not a single keystroke of
257 your work will be lost if the display device that you are using
258 crashes, or the network connection times out, or whatever. (I had an
259 extremely unstable X server for some time while I was developing these
260 patches, and running Emacs this way has saved me a number of M-x
261 recover-session invocations.)
263 I use the following two bash scripts to handle my Emacs sessions:
265 -------------------------------------------------------connect-emacs--
267 # Usage: connect-emacs <name> <args>...
269 # Connects to the Emacs instance called NAME. Starts up the instance
270 # if it is not already running. The rest of the arguments are passed
276 if [ -z "$name" ]; then
277 echo "Usage: connect_emacs <name> <args>..." >&2
280 preload-emacs "$name" wait
281 /usr/bin/emacsclient.emacs-multi-tty -s "$name" "$@"
282 ----------------------------------------------------------------------
284 -------------------------------------------------------preload-emacs--
286 # Usage: preload-emacs <name> [<waitp>]
288 # Preloads the Emacs instance called NAME in a detached screen
289 # session. Does nothing if the instance is already running. If WAITP
290 # is non-empty, the function waits until the server starts up and
291 # creates its socket; otherwise it returns immediately.
295 screendir="/var/run/screen/S-$USER"
296 serverdir="/tmp/emacs$UID"
297 emacs=/usr/bin/emacs-multi-tty # Or wherever you installed your multi-tty Emacs
299 if [ -z "$name" ]; then
300 echo "Usage: preload_emacs <name> [<waitp>]" >&2
304 if [ ! -e "$screendir"/*."$name" ]; then
305 if [ -e "$serverdir/$name" ]; then
306 # Delete leftover socket (for the wait option)
307 rm "$serverdir/$name"
309 screen -dmS "$name" "$emacs" -nw --eval "(setq server-name \"$name\")" -f server-start
311 if [ ! -z "$waitp" ]; then
312 while [ ! -e "$serverdir/$name" ]; do sleep 0.1; done
314 ----------------------------------------------------------------------
316 I have the following in my profile to have two instances automatically
317 preloaded for editing and email:
322 It is useful to set up short aliases for connect-emacs. I use the
325 alias edit="connect-emacs editor"
327 alias et="connect-emacs editor -t"
328 alias gnus="connect-emacs gnus"
340 For the NEWS file: (Needs much, much work)
342 ** Support for multiple terminal devices has been added.
344 *** You can specify a terminal device (`tty' parameter) and a terminal
345 type (`tty-type' parameter) to `make-terminal-frame'.
347 *** You can test for the presence of multiple terminal support by
348 testing for the `multi-tty' feature.
350 *** Emacsclient has been extended to support opening a new terminal
351 frame. Its behaviour has been changed to open a new Emacs frame by
352 default. Use the -c option to get the old behavior of opening
353 files in the currently selected Emacs frame.
355 *** A make-frame-on-tty function has been added to make it easier to
356 create frames on new terminals.
358 *** New functions: frame-tty-name, frame-tty-type, delete-tty,
359 terminal-local-value, set-terminal-local-value
361 terminal-id, terminal-parameters, terminal-parameter,
362 set-terminal-parameter
364 *** New variables: global-key-translation-map
366 *** The keymaps key-translation-map and function-key-map are now
369 ** Support for simultaneous graphical and terminal frames has been
372 *** The function `make-frame-on-display' now works during a terminal
373 session, and `make-frame-on-tty' works during a graphical session.
375 *** The `window-system' variable has been made frame-local.
377 *** The new `initial-window-system' variable contains the
378 `window-system' value for the first frame.
380 *** talk.el has been extended for multiple tty support.
385 (The rest of this file consists of my development notes and as such it
386 is probably not very interesting for anyone else.)
391 ** Report GTK multi-display problems to GTK maintainers. For extra
394 ** frames-on-display-list should also accept frames.
396 ** I smell something funny around pop_kboard's "deleted kboard" case.
397 Determine what are the circumstances of this case, and fix any
398 bug that comes to light.
400 ** Consider the `tty-type' frame parameter and the `display-tty-type'
401 function. They serve the exact same purpose. I think it may be
402 a good idea to eliminate one of them, preferably `tty-type'.
404 ** The handling of lisp/term/*.el, and frame creation in general, is a
405 big, big mess. How come the terminal-specific file is loaded by
406 tty-create-frame-with-faces? I don't think it is necessary to load
407 these files for each frame; once per terminal should be enough.
408 Update: lisp/term/*.el is not loaded repeatedly anymore, but
409 faces.el still needs to be cleaned up.
411 ** Fix frame-set-background-mode in this branch. It was recently
412 changed in CVS, and frame.el in multi-tty has not yet been adapted
413 for the changes. (It needs to look at
414 default-frame-background-mode.) (Update: maybe it is fixed now;
417 ** I think `(set-)terminal-local-value' and the terminal parameter
418 mechanism should be integrated into a single framework.
420 ** Add the following hooks: after-delete-frame-hook (for server.el,
421 instead of delete-frame-functions),
422 after-delete-terminal-functions, after-create-terminal-functions.
424 ** If the first key pressed on a new tty terminal is a function key,
425 it is not recognized correctly. May be related to the bug below.
427 ** Fix set-input-mode for multi-tty. It's a truly horrible interface;
428 what if we'd blow it up into several separate functions (with a
429 compatibility definition)?
431 ** BULK RENAME: The `display-' prefix of new Lisp-level functions
432 conflicts with stuff like `display-time-mode'. Use `device-'
433 or `terminal-' instead. I think I prefer `terminal-'.
435 It turns out that most of the offending Lisp functions were defined
436 in the trunk. Therefore, compatibility aliases should be defined
437 for the following names:
439 display-color-cells terminal-color-cells
440 display-color-p terminal-color-p
441 display-graphic-p terminal-graphic-p
442 display-grayscale-p terminal-grayscale-p
443 display-images-p terminal-images-p
444 display-mm-height terminal-mm-height
445 display-mm-width terminal-mm-width
446 display-mouse-p terminal-mouse-p
447 display-multi-font-p terminal-multi-font-p
448 display-multi-frame-p terminal-multi-frame-p
449 display-pixel-height terminal-pixel-height
450 display-pixel-width terminal-pixel-width
451 display-pixels-per-inch terminal-pixels-per-inch
452 display-planes terminal-planes
453 display-popup-menus-p terminal-popup-menus-p
454 display-save-under terminal-save-under
455 display-screens terminal-screens
456 display-supports-face-attributes-p terminal-supports-face-attributes-p
457 display-visual-class terminal-visual-class
458 framep-on-display framep-on-terminal
459 frames-on-display-list frames-on-terminal-list
461 The following functions were introduced in the multi-tty branch, and
462 can be renamed without aliases:
464 display-controlling-tty-p terminal-controlling-tty-p
465 display-list terminal-list
466 display-live-p terminal-live-p
467 display-name terminal-name
468 display-tty-type terminal-tty-type
469 frame-display terminal-of-frame
470 delete-display delete-terminal
472 ** The semantics of terminal-local variables are confusing; it is not
473 clear what binding is in effect in any given time. See if
474 current_kboard (or at least the terminal-local bindings exported to
475 Lisp) might be changed to be tied to the selected frame instead.
476 Currently, `function-key-map' and `key-translation-map' may be
477 accessed reliably only using the hackish
478 `(set-)terminal-local-value' functions.
480 Perhaps there should be a difference between `last-command' &co.
481 and these more conventional configuration variables.
482 (E.g. `symbol-value' would use current_kboard to access
483 `last-command', but SELECTED_FRAME()->display->kboard to get the
484 value of `function-key-map'.
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 ** normal-erase-is-backspace-mode in simple.el needs to be updated for
537 multi-tty (rep. by Dan Waber).
539 ** Hunt down display-related functions in frame.el and extend them all
540 to accept display ids.
542 ** rif->flush_display_optional (NULL) calls should be replaced by a
545 ** The set-locale-environment hack (adding the DISPLAY option) should
546 be replaced with a clean design.
548 ** standard-display-table should be display-local.
549 standard-display-european should be display-local.
551 ** With iswitchb-default-method set to 'always-frame, only frames on
552 the current display should be considered. This might involve
553 extending `get-buffer-window'.
555 ** Have a look at Vlocale_coding_system. Seems like it would be a
556 tedious job to localize it, although most references use it for
557 interfacing with libc and are therefore OK with the global
560 Exceptions found so far: x-select-text and
561 x-cut-buffer-or-selection-value.
563 ** Have a look at fatal_error_hook.
565 ** Have a look at set_frame_matrix_frame.
567 ** Check if we got term-setup-hook right.
569 ** I think tip_frame should be display-local.
571 ** Check display reference count handling in x_create_tip_frame.
573 ** make-frame does not correctly handle extra parameters in its
576 (frame-parameter (make-frame (list (cons 'foobar 42))) 'foobar)
579 (This is likely an error in the CVS trunk.)
581 ** Dan Nicolaescu suggests that -nw should be added as an alias for -t
582 in emacsclient. Good idea. (Alas, implementing this is not
583 trivial, getopt_long does not seem to support two-letter ``short''
584 options. Patches are welcome.) :-)
586 ** Mark Plaksin suggests that emacsclient should accept the same
587 X-related command-line arguments as Emacs. Most of the X-related
588 argument-handling is done in Lisp, so this should be quite easy to
591 ** Gergely Nagy suggests that C-x # should only kill the current
592 frame, not any other emacsclient frame that may have the same file
593 opened for editing. I think I agree with him.
595 ** Very strange bug: visible-bell does not work on secondary
596 terminals in xterm and konsole. The screen does flicker a bit,
597 but it's so quick it isn't noticable.
599 ** Move baud_rate to struct display.
601 ** Implement support for starting an interactive Emacs session without
602 an initial frame. (The user would connect to it and open frames
603 later, with emacsclient.)
605 ** Fix Mac support (I can't do this entirely myself). Note that the
606 current state of Mac-specific source files in the multi-tty tree
607 are not useful; before starting work on Mac support, revert to
608 pristine, pre-multi-tty versions.
610 ** Fix W32 support (I can't do this entirely myself). Note that the
611 current state of W32-specific source files in the multi-tty tree
612 are not useful; before starting work on W32 support, revert to
613 pristine, pre-multi-tty versions.
615 ** Fix DOS support (I can't do this entirely myself). Note that the
616 current state of DOS-specific source files in the multi-tty tree
617 are not useful; before starting work on DOS support, revert to
618 pristine, pre-multi-tty versions.
620 ** Do a grep on XXX and ?? for more issues.
622 ** Understand Emacs's low-level input system (it's black magic) :-)
623 What exactly does interrupt_input do? I tried to disable it for
624 raw secondary tty support, but it does not seem to do anything
625 useful. (Update: Look again. X unconditionally enables this, maybe
626 that's why raw terminal support is broken again. I really do need
627 to understand input.)
629 ** Fix stuff_char for multi-tty. Doesn't seem to be of high priority.
634 (ex-TODO items with explanations.)
636 -- Introduce a new struct for terminal devices.
638 (Done, see struct tty_output. The list of members is not yet
641 -- Change the bootstrap procedure to initialize tty_list.
643 (Done, but needs review.)
645 -- Change make-terminal-frame to support specifying another tty.
647 (Done, new frame parameters: `tty' and `tty-type'.)
649 -- Implement support for reading from multiple terminals.
651 (Done, read_avail_input tries to read from each terminal, until one
652 succeeds. MULTI_KBOARD is not used. Secondary terminals don't send
655 (Update: They do, now.)
657 (Update2: After enabling X, they don't.)
659 -- other-frame should cycle through the frames on the `current'
662 (Done, by trivially modifiying next_frame and prev_frame.)
664 -- Support different terminal sizes.
668 -- Make sure terminal resizes are handled gracefully. (Could be
671 (Done. We don't get automatic SIGWINCH for additional ttys,
674 -- Extend emacsclient to automatically open a new tty when it connects
677 (Done. It's an ugly hack, needs more work.)
679 -- Redisplay must refresh the topmost frame on *all* terminals, not
680 just the initial terminal.
682 (Done, but introduced an ugly redisplay problems. Ugh.)
684 -- Fix redisplay problems.
686 (Done; it turned out that the entire Wcm structure must be moved
687 inside tty_output. Why didn't I catch this earlier?)
689 -- Provide a way for emacsclient to tell Emacs that the tty has been
692 (Done, simply forward the SIGWINCH signal.)
694 -- Each keypress should automatically select the frame corresponding
695 to the terminal that it was coming from. This means that Emacs
696 must know from which terminal the last keyboard event came from.
698 (Done, it was quite simple, the input event system already
699 supported multiple frames.)
701 -- Fix SIGIO issue with secondary terminals.
703 (Done, emacsclient signals Emacs after writing to the proxy pseudo
704 terminal. Note that this means that multi-tty does not work with
707 (Update: This is bullshit. There is a read_input_waiting function,
708 extend that somehow.)
710 (Update of update: The first update was not right either, extending
711 read_input_waiting was not necessary. Secondary ttys do seem to
712 send signals on input.)
714 (Update^3: Not any more.)
716 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
717 parameters from the currently selected terminal before the global
722 -- Put all cached terminal escape sequences into struct tty_output.
723 Currently, they are still stored in global variables, so we don't
724 really support multiple terminal types.
726 (Done. It was not fun.)
728 -- Implement sane error handling after initialization. (Currently
729 emacs exits if you specify a bad terminal type.) The helpful error
730 messages must still be provided when Emacs starts.
734 -- Implement terminal deletion, i.e., deleting local frames, closing
735 the tty device and restoring its previous state without exiting
738 (Done, but at the moment only called when an error happens during
739 initialization. There is a memory corruption error around this
740 somewhere.) (Update: now it is fully enabled.)
742 -- Implement automatic deletion of terminals when the last frame on
743 that terminal is closed.
747 -- Restore tty screen after closing the terminal.
749 (Done, we do the same as Emacs 21.2 for all terminals.)
751 -- 'TERM=dumb src/emacs' does not restore the terminal state.
755 -- C-g should work on secondary terminals.
757 (Done, but the binding is not configurable.)
759 -- Deal with SIGHUP in Emacs and in emacsclient. (After this, the
760 server-frames may be removed from server.el.)
762 (Done, nothing to do. It seems that Emacs does not receive SIGHUP
763 from secondary ttys, which is actually a good thing.) (Update: I
764 think it would be a bad idea to remove server-frames.)
766 -- Change emacsclient/server.el to support the -t argument better,
767 i.e. automatically close the socket when the frame is closed.
769 (Seems to be working OK.)
771 -- Fix mysterious memory corruption error with tty deletion. To
772 trigger it, try the following shell command:
774 while true; do TERM=no-such-terminal-definition emacsclient -h; done
776 Emacs usually dumps core after a few dozen iterations. (The bug
777 seems to be related to the xfreeing or bzeroing of
778 tty_output.Wcm. Maybe there are outside references to struct Wcm?
779 Why were these vars collected into a struct before multi-tty
782 (Done. Whew. It turned out that the problem had nothing to do
783 with hypothetical external references to Wcm, or any other
784 tty_output component; it was simply that delete_tty closed the
785 filehandles of secondary ttys twice, resulting in fclose doubly
786 freeing memory. Utterly trivial matter. I love the C's memory
787 management, it puts hair on your chest.)
789 -- Support raw secondary terminals. (Note that SIGIO works only on
790 the controlling terminal.) Hint: extend read_input_waiting for
791 multiple ttys and hopefully this will be fixed.
793 (Done, it seems to have been working already for some time. It
794 seems F_SETOWN does work, after all. Not sure what made it fail
795 earlier, but it seems to be fixed (there were several changes
796 around request_sigio, maybe one of them did it).
797 read_input_waiting is only used in sys_select, don't change
798 it.) (Update: After adding X support, it's broken again.)
799 (Update^2: No it isn't.) :-)
801 -- Find out why does Emacs abort when it wants to close its
802 controlling tty. Hint: chan_process[] array. Hey, maybe
803 noninterrupt-IO would work, too? Update: no, there is no process
806 (Done. Added add/delete_keyboard_wait_descriptor to
807 term_init/delete_tty. The hint was right, in a way.)
809 -- Issue with SIGIO: it needs to be disabled during redisplay. See if
810 fcntl kernel behaviour could be emulated by emacsclient.
812 (Done. Simply disabled the SIGIO emulation hack in emacsclient.)
813 (Update: it was added back.) (Update^2: and removed again.)
815 -- server.el: There are issues with saving files in buffers of closed
816 clients. Try editing a file with emacsclient -f, and (without
817 saving it) do a delete-frame. The frame is closed without
818 question, and a surprising confirmation prompt appears in another
821 (Done. delete-frame now asks for confirmation if it still has
822 pending buffers, and modified buffers don't seem to be deleted.)
824 -- emacsclient.el, server.el: Handle eval or file open errors when
829 -- Make parts of struct tty_output accessible from Lisp. The device
830 name and the type is sufficient.
832 (Done, see frame-tty-name and frame-tty-type.)
834 -- Export delete_tty to the Lisp environment, for emacsclient.
836 (Done, see delete-tty.)
838 -- Get rid of the accessor macros in termchar.h, or define macros for
843 -- Move device-specific parameters (like costs) commonly used by
844 device backends to a common, device-dependent structure.
846 (Done. See struct display_method in termhooks.h.)
850 (Done. Well, it seems to be working.)
852 -- Allow simultaneous X and tty frames. (Handling input could be
853 tricky. Or maybe not.)
855 (Done. Allowed, that is. It is currently extremely unstable, to
856 the point of being unusable. The rif variable causes constant
857 core dumps. Handling input is indeed tricky.)
859 -- Rewrite multi-tty input in terms of MULTI_KBOARD.
861 (Done. In fact, there was no need to rewrite anything, I just
862 added a kboard member to tty_display_info, and initialized the
863 frame's kboard from there.)
865 -- Fix rif issue with X-tty combo sessions. IMHO the best thing to do
866 is to get rid of that global variable (and use the value value in
867 display_method, which is guaranteed to be correct).
869 (Done, did exactly that. Core dumps during combo sessions became
870 much rarer. In fact, I have not yet met a single one.)
872 -- Add multi-tty support to talk.el.
876 -- Clean up the source of emacsclient. It is a mess.
878 (Done, eliminated stupid proxy-pty kludge.)
880 -- Fix faces on tty frames during X-tty combo sessions. There is an
881 init_frame_faces call in init_sys_modes, see if there is a problem
884 (Done, there was a stupid mistake in
885 Ftty_supports_face_attributes_p. Colors are broken, though.)
887 -- C-x 5 2, C-x 5 o, C-x 5 0 on an emacsclient frame unexpectedly
888 exits emacsclient. This is a result of trying to be clever with
889 delete-frame-functions.
891 (Fixed, added delete-tty-after-functions, and changed server.el to
894 -- Something with (maybe) multi-keyboard support broke function keys
895 and arrows on ttys during X+tty combo sessions. Debug this.
897 (I can't reproduce it, maybe the terminal type was wrong.)
899 -- Fix input from raw ttys (again).
901 (Now it seems to work all right.)
903 -- During an X-tty combo session, a (message "Hello") from a tty frame
904 goes to the X frame. Fix this.
906 (Done. There was a safeguard against writing to the initial
907 terminal frame during bootstrap which prevented echo_area_display
908 from working correctly on a tty frame during a combo session.)
910 -- If there are no frames on its controlling terminal, Emacs should
911 exit if the user presses C-c there.
913 (Done, as far as possible. See the SIGTERM comment in
914 interrupt_signal on why this seems to be impossible to solve this
917 -- During an X session, Emacs seems to read from stdin. Also, Emacs
918 fails to start without a controlling tty.
920 (Fixed by replacing the troublesome termcap display with a dummy
921 bootstrap display during bootstrap.
923 -- Do tty output through struct display, like graphical display
928 -- Define an output_initial value for output_method for the initial
929 frame that is dumped with Emacs. Checking for this frame (e.g. in
930 cmd_error_internal) is ugly.
932 (Done, broking interactive temacs.)
934 -- The command `emacsclient -t -e '(delete-frame)'' fails to exit.
938 -- frame-creation-function should always create a frame that is on the
939 same display as the selected frame. Maybe frame-creation-function
940 should simply be removed and make-frame changed to do the right
943 (Done, with a nice hack. frame-creation-function is now frame-local.)
945 -- Fix C-g on raw ttys.
947 (Done. I disabled the interrupt/quit keys on all secondary
948 terminals, so Emacs sees C-g as normal input. This looks like an
949 overkill, because emacsclient has extra code to pass SIGINT to
950 Emacs, so C-g should remain the interrupt/quit key on emacsclient
951 frames. See the next entry why implementing this distinction would
954 -- Make sure C-g goes to the right frame with ttys. This is hard, as
955 SIGINT doesn't have a tty parameter. :-(
957 (Done, the previous change fixes this as a pleasant side effect.)
959 -- I have seen a case when Emacs with multiple ttys fell into a loop
960 eating 100% of CPU time. Strace showed this loop:
963 kill(30284, SIGIO) = 0
964 --- SIGIO (I/O possible) @ 0 (0) ---
965 ioctl(6, FIONREAD, [0]) = -1 EIO (Input/output error)
966 ioctl(5, FIONREAD, [0]) = -1 EIO (Input/output error)
967 ioctl(0, FIONREAD, [0]) = 0
968 sigreturn() = ? (mask now [])
969 gettimeofday({1072842297, 747760}, NULL) = 0
970 gettimeofday({1072842297, 747806}, NULL) = 0
971 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
972 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
973 gettimeofday({1072842297, 748245}, NULL) = 0
975 I have seen something similar with a single X frame, but have not
976 been able to reproduce it for debugging.
978 Update: This may have been caused by checking for nread != 0
979 instead of nread > 0 after calling read_socket_hook in
982 (Fixed. This was caused by unconditionally including stdin in
983 input_wait_mask in init_process. The select call in
984 wait_reading_process_input always returned immediately, indicating
985 that there is pending input from stdin, which nobody read.
987 Note that the above strace output seems to be an unrelated but
988 similar bug. I think that is now fixed.)
990 -- Exiting Emacs while there are emacsclient frames doesn't restore the
991 ttys to their default states.
993 (This seems to be fixed by some previous change.)
995 -- Allow opening an X session after -nw.
999 -- Fix color handling during tty+X combo sessions. (It seems that tty
1000 sessions automatically convert the face colors to terminal colors
1001 when the face is loaded. This conversion must happen instead on
1002 the fly in write_glyphs, which might be problematic, as color
1003 approximation is currently done in lisp (term/tty-colors.el).)
1004 (Update: hm, colors seem to work fine if I start emacs with -nw and
1005 then create an X frame. Maybe it's just a small buglet somewhere.)
1007 (Seems to be fixed. The problem was in startup.el, it did not
1008 initialize tty colors when the initial window system was
1011 -- emacs -nw --eval '(y-or-n-p "Foobar")' segfaults. (Reported by
1014 (Fixed, there was a keyboard initialization problem.)
1016 -- Fix interactive use of temacs. There are face-related SEGVs, most
1017 likely because of changes in realize_default_face, realize_face.
1021 -- Don't exit Emacs when the last X connection fails during a
1022 multi-display session.
1026 -- Dan Nicolaescu noticed that starting emacsclient on the same
1027 terminal device that is the controlling tty of the Emacs process
1028 gives unexpected results.
1032 -- Istvan Marko reported that Emacs hang on ttys if it was started
1033 from a shell script.
1035 (Fixed. There was a bug in the multi-tty version of
1036 narrow_foreground_group. tcsetpgrp blocks if it is called from a
1037 process that is not in the same process group as the tty.)
1039 -- emacsclient -t from an Emacs term buffer does not work, complains
1040 about face problems. This can even lock up Emacs (if the recursive
1041 frame sets single_kboard). Update: the face problems are caused by
1042 bugs in term.el, not in multi-tty. The lockup is caused by
1043 single_kboard mode, and is not easily resolvable. The best thing to
1044 do is to simply refuse to create a tty frame of type `eterm'.
1046 (Fixed, changed emacsclient to check for TERM=eterm. The face
1047 complaints seem to be caused by bugs in term.el; they are not
1048 related to multi-tty.)
1050 -- Find out the best way to support suspending Emacs with multiple
1051 ttys. My guess: disable it on the controlling tty, but from other
1052 ttys pass it on to emacsclient somehow. (It is (I hope) trivial to
1053 extend emacsclient to handle suspend/resume. A `kill -STOP' almost
1056 (Done. I needed to play with signal handling and the server
1057 protocol a bit to make emacsclient behave as a normal UNIX program
1058 wrt foreground/background process groups.)
1060 -- There is a flicker during the startup of `emacs -nw'; it's as if
1061 the terminal is initialized, reset and then initialialized again.
1062 Debug this. (Hint: narrow_foreground_group is called twice during
1067 -- Robert Chassell has found serious copy-paste bugs with the
1068 multi-tty branch. There seem to be redisplay bugs while copying
1069 from X to a terminal frame. Copying accented characters do not
1072 (Patch-124 should fix this, by changing the interprogram-*-function
1073 variables to be frame-local, as suggested by Mark Plaksin
1074 (thanks!). I think that the redisplay bugs are in fact not bugs,
1075 but delays caused by single_kboard --> perhaps MULTI_KBOARD should
1078 -- frame-creation-function was removed, which might be a bad idea.
1079 Think up a compatible solution.
1081 (It was an internal interface that may be changed when necessary.)
1083 -- Change Lisp code not to (getenv "TERM"); use the `tty-type' frame
1084 parameter or the frame-tty-type function instead. (M-x tags-search
1085 "TERM" helps with this.) Update: Actually, all getenv invocations
1086 should be checked for multi-tty compatibility, and an interface
1087 must be implemented to get the remote client's environment.
1089 (Done. Only getenv calls in lisp/term/*.el were changed; other
1090 calls should be mostly left as they are.)
1092 -- Add an elaborate mechanism for display-local variables. (There are
1093 already a few of these; search for `terminal-local' in the Elisp
1096 (Not needed. Display-local variables could be emulated by
1097 frame-local variables.)
1099 -- Emacs assumes that all terminal frames have the same locale
1100 settings as Emacs itself. This may lead to bogus results in a
1101 multi-locale setup. (E.g., while logging in from a remote client
1102 with a different locale.)
1103 (Update after new bugreport by Friedrich Delgado Friedrichs:
1104 (at least) the structs terminal_coding and keyboard_coding in
1105 coding.c must be moved to struct display, and the Lisp interface
1106 [set-]keyboard-coding-system must be adapted for the change.)
1108 (Fixed. Emacs now uses the locale settings as seen by the
1109 emacsclient process for server tty frames.)
1110 (Update: Not really; Vlocale_coding_system is still global.)
1112 -- Make `struct display' accessible to Lisp programs. Accessor functions:
1114 (displayp OBJECT): Returns t if OBJECT is a display.
1115 => Implemented as display-live-p.
1117 (display-list): Returns list of currently active displays.
1120 (selected-display): Returns the display object of the selected frame.
1121 => Not strictly necessary, but implemented anyway.
1123 (frame-display FRAME): Returns the display object of FRAME.
1126 (display-frames DISPLAY): Returns a list of frames on DISPLAY.
1127 => Already implemented, see frames-on-display-list.
1129 (display-type DISPLAY): Returns the type of DISPLAY, as a
1130 symbol. (See `framep'.)
1131 => Implemented as display-live-p.
1133 (display-device DISPLAY): Returns the name of the device that
1134 DISPLAY uses, as a string. (E.g: "/dev/pts/16", or
1136 => Implemented as display-name.
1140 See next issue why this is necessary.
1142 (Update: The consensus on emacs-devel seems to be to do this via
1143 integer identifiers. That's fine by me.)
1147 -- The following needs to be supported:
1154 The cleanest way to solve this is to allow multiple displays on the
1155 same terminal device; each new emacsclient process should create
1156 its own display. As displays are currently identified by their
1157 device names, this is not possible until struct display becomes
1158 accessible as a Lisp-level object.
1162 -- Miles Bader suggests that C-x C-c on an emacsclient frame should
1163 only close the frame, not exit the entire Emacs session. Update:
1164 see above for a function that does this. Maybe this should be the
1167 (Done. This is the new default. No complaints so far.)
1169 -- Clean up the frame-local variable system. I think it's ugly and
1170 error-prone. But maybe I just haven't yet fully understood it.
1172 (Nothing to do. It doesn't seem ugly any more. It's rather clever.)
1174 -- Support multiple character locales. A version of
1175 `set-locale-environment' needs to be written for setting up
1176 display-local settings on ttys. I think calling
1177 set-display-table-and-terminal-coding-system and
1178 set-keyboard-coding-system would be enough. The language
1179 environment itself should remain a global setting.
1181 (Done, by an ugly hack.)
1183 -- The terminal customization files in term/*.el tend to change global
1184 parameters, which may confuse Emacs with multiple displays. Change
1185 them to tweak only frame-local settings, if possible. (They tend
1186 to call define-key to set function key sequences a lot.)
1188 (Done, by making `function-key-map' terminal-local (i.e., part of
1189 struct kboard). This has probably covered all the remaining problems.)
1191 -- Make `function-key-map' and `key-translation-map' terminal-local.
1195 -- Implement `terminal-local-value' and `set-terminal-local-value' to
1196 allow deterministic access to terminal local bindings. The
1197 encode-kb package can not set up `key-translation-map' without
1198 these. The terminal-local bindings seem to be independent of what
1203 -- xt-mouse.el needs to be adapted for multi-tty. It currently
1204 signals an error on kill-emacs under X, which prevents the user
1205 from exiting Emacs. (Reported by Mnemonikk on freenode.)
1210 -- Having {reset,init}_all_sys_modes in set-input-mode breaks arrow
1211 keys on non-selected terminals under screen, and sometimes on other
1212 terminal types as well. The other function keys continue to work
1213 fine. Sometimes faces on these screens become garbled.
1215 This only seems to affect displays that are of the same terminfo
1216 type as the selected one. Interestingly, in screen Emacs normally
1217 reports the up arrow key as `M-o A', but after the above SNAFU, it
1218 complains about `M-[ a'. UNIX ttys are a complete mystery to me,
1219 but it seems the reset-reinitialize cycle somehow leaves the
1220 non-selected terminals in a different state than usual. I have no
1221 idea how this could happen.
1223 Currently set-input-mode resets only the currently selected
1224 terminal, which seems to somehow work around the problem.
1228 Dan Nicolaescu <dann@ics.uci.edu> writes:
1229 > Some terminals have 2 modes for cursor keys: Application Mode where
1230 > the cursor keys transmit the codes defined in the terminfo entry, and
1231 > Cursor mode. Applications have to send the smkx and rmkx terminfo
1232 > strings to switch between the 2 modes. So Emacs (and emacsclient) have
1233 > to send smkx when initializing and rmkx when quitting (or on
1236 (I think patch-370 fixed this.)
1238 -- This long-standing bug (first reported by Han Boetes) seems to come
1239 and go all the time. It is time to track it down and fix it.
1244 # From another xterm:
1245 emacsclient -e '(y-or-n-p "Do you want me to crash? ")'
1246 # Notice how the answer ends up in the *scratch* buffer
1250 (Fixed in patch-414 after detailed analysis by Kalle Olavi Niemitalo.)
1252 ;;; arch-tag: 8da1619e-2e79-41a8-9ac9-a0485daad17d