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 at mbf dot ocn dot ne dot jp>
35 Vincent Bernat <bernat at luffy dot cx>
36 Han Boetes <han at mijncomputer dot nl>
37 Robert J. Chassell <bob at rattlesnake dot com>
38 Romain Francoise <romain at orebokech dot com>
39 Ami Fischman <ami at fischman dot org>
40 Friedrich Delgado Friedrichs <friedel at nomaden dot org>
41 IRIE Tetsuya <irie at t dot email dot ne dot jp>
42 Yoshiaki Kasahara <kasahara at nc dot kyushu-u dot ac dot jp>
43 Jurej Kubelka <Juraj dot Kubelka at email dot cz>
44 David Lichteblau <david at lichteblau dot com>
45 Xavier Mallard <zedek at gnu-rox dot org>
46 Istvan Marko <mi-mtty at kismala dot com>
47 Ted Morse <morse at ciholas dot com>
48 Dan Nicolaescu <dann at ics dot uci dot edu>
49 Gergely Nagy <algernon at debian dot org>
50 Mark Plaksin <happy at mcplaksin dot org>
51 Francisco Borges <borges at let dot rug dot nl>
52 Frank Ruell <stoerte at dreamwarrior dot net>
53 Dan Waber <dwaber at logolalia dot com>
56 Richard Stallman was kind enough to review an earlier version of my
63 The multi-tty mailing list (discussion & bug reports):
65 Address: multi-tty@lists.fnord.hu
66 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty
67 Archive: http://news.gmane.org/gmane.emacs.multi-tty/
69 Commit notifications (read-only):
71 Address: multi-tty-commits@lists.fnord.hu
72 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty-commits
78 The branch is now very stable and almost full-featured. All of the
79 major problems have been fixed, only a few minor issues remain. (It
80 still needs to be ported to Windows/Mac/DOS, though.) Both multiple
81 tty device support and simultaneous X and tty frame support works
82 fine. Emacsclient has been extended to support opening new tty and X
83 frames. It has been changed open new Emacs frames by default.
85 The multi-tty branch has been scheduled for inclusion in the next
86 major release of Emacs (version 23). I expect the merge into the
87 development trunk to occur sometime during next year (2005), after the
88 merge of the Unicode branch.
90 Tested on GNU/Linux, Solaris 8, FreeBSD and OpenBSD. Please let me
91 know if you succeed or fail to use it on other platforms---I'll have a
92 few tricky test cases for you.
96 * GTK support. If you compile your Emacs with the GTK
97 toolkit, some functionality of multi-tty will be lost.
98 In particular, closing an X display causes a crash.
99 Current releases of GTK have limitations and bugs that
100 prevent full-blown multi-display support in Emacs. Use the
101 Lucid toolkit if you want to see a complete feature set.
103 * The single-kboard mode.
105 If your multi-tty Emacs session seems to be frozen, you
106 probably have a recursive editing session or a pending
107 minibuffer prompt (which is a kind of recursive editing) on
108 another display. To unfreeze your session, switch to that
109 display and complete the recursive edit, for example by
110 pressing C-] (`abort-recursive-edit').
112 I am sorry to say that currently there is no way to break
113 out of this "single-kboard mode" from a frozen display. If
114 you are unable to switch to the display that locks the
115 others (for example because it is on a remote computer),
116 then you can use emacsclient to break out of all recursive
119 emacsclient -e '(top-level)'
121 Note that this (perhaps) unintuitive behaviour is by design.
122 Single-kboard mode is required because of an intrinsic Emacs
123 limitation that is very hard to eliminate. (This limitation
124 is related to the single-threaded nature of Emacs.)
126 I plan to implement better user notification and support for
127 breaking out of single-kboard mode from locked displays.
129 * Mac, Windows and DOS support is broken, doesn't even
130 compile. Multiple display support will probably not provide
131 new Emacs features on these systems, but the multi-tty
132 branch changed a few low-level interfaces, and the
133 system-dependent source files need to be adapted
134 accordingly. The changes are mostly trivial, so almost
135 anyone can help, if only by compiling the branch and
136 reporting the compiler errors. (It is not worth to do this
140 HOW TO GET THE BRANCH
141 ---------------------
143 The branch uses GNU Arch (http://www.gnuarch.org) for version control.
145 Retrieving the latest version of the branch:
147 tla register-archive -f http://lorentey.hu/arch/2004/
148 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
150 This incantation uses my private archive mirror that is hosted on a
151 relatively low-bandwidth site; if you are outside Hungary, you will
152 probably want to you use my secondary mirror: (Note that the -f option
153 will overwrite the archive location if you have previously registered
156 tla register-archive -f http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
157 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
159 http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
161 The Arch supermirror provides mirroring services for all public Arch
162 repositories. We have a mirror there, too, if you prefer.
164 tla register-archive -f http://mirrors.sourcecontrol.net/lorentey%40elte.hu--2004
165 tla get lorentey@elte.hu--2004/emacs--multi-tty <directory>
167 My GPG key id is 0FB27A3F; it is available from
168 hkp://wwwkeys.eu.pgp.net/, or from my homepage at
169 http://lorentey.hu/rolam/gpg.html)
171 Don't worry if the above checkout takes a few minutes to complete;
172 once you have a source tree, updating it to the latest revision will
173 be _much_ faster. Use the following command for the update:
177 You can find more information about Arch on http://wiki.gnuarch.org/.
178 It's a wonderful source control system, I highly recommend it.
180 If you don't have tla, the branch has a homepage from which you can
181 download conventional patches against Emacs CVS HEAD:
183 http://lorentey.hu/project/emacs.html
188 If you run Debian, or a distribution based on Debian, you are welcome
189 to use our binary packages; put these lines in your /etc/apt/sources.list:
192 deb http://aszt.inf.elte.hu/~lorentey/mirror/apt ./
193 deb-src http://aszt.inf.elte.hu/~lorentey/mirror/apt ./
195 Note that these packages are intended solely to provide an easy way to
196 test the new multi-tty features. They are not to be taken as Emacs
197 releases, and it's a mistake to expect robust operation or any kind of
198 timely support for them. Do not install them, unless you'd like to
199 have your editor crash on you.
205 The multi-tty branch is compiled the same way as Emacs itself:
207 make maintainer-clean # (If you have compiled Emacs before)
209 ./configure <your favourite options>
213 If you have strange compilation errors, they may be caused by old
214 *.elc files that are left over from an earlier bootstrap. The `make
215 maintainer-clean' target deletes them, so it is a good idea to run
216 that before reporting a bug. (Emacs requires a clean recompilation
217 after certain kinds of source code changes.)
222 To test the multi-tty branch, start up the Emacs server with the
228 and then (from a shell prompt on another terminal) start emacsclient
230 emacsclient -t /optional/file/names... (for a tty frame)
231 emacsclient /optional/file/names... (for an X frame)
233 (Make sure both emacs and emacsclient are multi-tty versions.)
234 You'll hopefully have two fully working, independent frames on
235 separate terminals. The new frame is closed automatically when you
236 finish editing the specified files (C-x #), but delete-frame (C-x 5 0)
237 also works. Of course, you can create frames on more than two tty
240 Creating new frames on the same tty with C-x 5 2 works, and they
241 behave the same way as in previous Emacs versions. If you exit emacs,
242 all terminals should be restored to their previous states.
244 This is work in progress, and probably full of bugs. It is a good
245 idea to run emacs from gdb, so that you'll have a live instance to
246 debug if something goes wrong. Please send me your bug reports on our
247 mailing list: multi-tty@lists.fnord.hu
252 I think the best way to use the new Emacs is to have it running inside
253 a disconnected GNU screen session, and always use emacsclient for
254 normal work. One advantage of this is that not a single keystroke of
255 your work will be lost if the display device that you are using
256 crashes, or the network connection times out, or whatever. (I had an
257 extremely unstable X server for some time while I was developing these
258 patches, and running Emacs this way has saved me a number of M-x
259 recover-session invocations.)
261 I use the following two bash scripts to handle my Emacs sessions:
263 -------------------------------------------------------connect-emacs--
265 # Usage: connect-emacs <name> <args>...
267 # Connects to the Emacs instance called NAME. Starts up the instance
268 # if it is not already running. The rest of the arguments are passed
274 if [ -z "$name" ]; then
275 echo "Usage: connect_emacs <name> <args>..." >&2
278 preload-emacs "$name" wait
279 /usr/bin/emacsclient.emacs-multi-tty -s "$name" "$@"
280 ----------------------------------------------------------------------
282 -------------------------------------------------------preload-emacs--
284 # Usage: preload-emacs <name> [<waitp>]
286 # Preloads the Emacs instance called NAME in a detached screen
287 # session. Does nothing if the instance is already running. If WAITP
288 # is non-empty, the function waits until the server starts up and
289 # creates its socket; otherwise it returns immediately.
293 screendir="/var/run/screen/S-$USER"
294 serverdir="/tmp/emacs$UID"
295 emacs=/usr/bin/emacs-multi-tty # Or wherever you installed your multi-tty Emacs
297 if [ -z "$name" ]; then
298 echo "Usage: preload_emacs <name> [<waitp>]" >&2
302 if [ ! -e "$screendir"/*."$name" ]; then
303 if [ -e "$serverdir/$name" ]; then
304 # Delete leftover socket (for the wait option)
305 rm "$serverdir/$name"
307 screen -dmS "$name" "$emacs" -nw --eval "(setq server-name \"$name\")" -f server-start
309 if [ ! -z "$waitp" ]; then
310 while [ ! -e "$serverdir/$name" ]; do sleep 0.1; done
312 ----------------------------------------------------------------------
314 I have the following in my profile to have two instances automatically
315 preloaded for editing and email:
320 It is useful to set up short aliases for connect-emacs. I use the
323 alias edit="connect-emacs editor"
325 alias et="connect-emacs editor -t"
326 alias gnus="connect-emacs gnus"
338 For the NEWS file: (Needs much, much work)
340 ** Support for multiple terminal devices has been added.
342 *** You can specify a terminal device (`tty' parameter) and a terminal
343 type (`tty-type' parameter) to `make-terminal-frame'.
345 *** You can test for the presence of multiple terminal support by
346 testing for the `multi-tty' feature.
348 *** Emacsclient has been extended to support opening a new terminal
349 frame. Its behaviour has been changed to open a new Emacs frame by
350 default. Use the -c option to get the old behavior of opening
351 files in the currently selected Emacs frame.
353 *** A make-frame-on-tty function has been added to make it easier to
354 create frames on new terminals.
356 *** New functions: frame-tty-name, frame-tty-type, delete-tty,
357 terminal-local-value, set-terminal-local-value
359 terminal-id, terminal-parameters, terminal-parameter,
360 set-terminal-parameter
362 *** New variables: global-key-translation-map
364 *** The keymaps key-translation-map and function-key-map are now
367 ** Support for simultaneous graphical and terminal frames has been
370 *** The function `make-frame-on-display' now works during a terminal
371 session, and `make-frame-on-tty' works during a graphical session.
373 *** The `window-system' variable has been made frame-local.
375 *** The new `initial-window-system' variable contains the
376 `window-system' value for the first frame.
378 *** talk.el has been extended for multiple tty support.
383 (The rest of this file consists of my development notes and as such it
384 is probably not very interesting for anyone else.)
389 ** This long-standing bug (first reported by Han Boetes) seems to come
390 and go all the time. It is time to track it down and fix it.
395 # From another xterm:
396 emacsclient -e '(y-or-n-p "Do you want me to crash? ")'
397 # Notice how the answer ends up in the *scratch* buffer
401 ** frames-on-display-list should also accept frames.
403 ** Consider the `tty-type' frame parameter and the `display-tty-type'
404 function. They serve the exact same purpose. I think it may be
405 a good idea to eliminate one of them, preferably `tty-type'.
407 ** The handling of lisp/term/*.el, and frame creation in general, is a
408 big, big mess. How come the terminal-specific file is loaded by
409 tty-create-frame-with-faces? I don't think it is necessary to load
410 these files for each frame; once per terminal should be enough.
411 Update: lisp/term/*.el is not loaded repeatedly anymore, but
412 faces.el still needs to be cleaned up.
414 ** Fix frame-set-background-mode in this branch. It was recently
415 changed in CVS, and frame.el in multi-tty has not yet been adapted
416 for the changes. (It needs to look at
417 default-frame-background-mode.) (Update: maybe it is fixed now;
420 ** I think `(set-)terminal-local-value' and the terminal parameter
421 mechanism should be integrated into a single framework.
423 ** Add the following hooks: after-delete-frame-hook (for server.el,
424 instead of delete-frame-functions),
425 after-delete-terminal-functions, after-create-terminal-functions.
427 ** If the first key pressed on a new tty terminal is a function key,
428 it is not recognized correctly. May be related to the bug below.
430 ** Having {reset,init}_all_sys_modes in set-input-mode breaks arrow
431 keys on non-selected terminals under screen, and sometimes on other
432 terminal types as well. The other function keys continue to work
433 fine. Sometimes faces on these screens become garbled.
435 This only seems to affect displays that are of the same terminfo
436 type as the selected one. Interestingly, in screen Emacs normally
437 reports the up arrow key as `M-o A', but after the above SNAFU, it
438 complains about `M-[ a'. UNIX ttys are a complete mystery to me,
439 but it seems the reset-reinitialize cycle somehow leaves the
440 non-selected terminals in a different state than usual. I have no
441 idea how this could happen.
443 Currently set-input-mode resets only the currently selected
444 terminal, which seems to somehow work around the problem.
446 ** Fix set-input-mode for multi-tty. It's a truly horrible interface;
447 what if we'd blow it up into several separate functions (with a
448 compatibility definition)?
450 ** BULK RENAME: The `display-' prefix of new Lisp-level functions
451 conflicts with stuff like `display-time-mode'. Use `device-'
452 or `terminal-' instead. I think I prefer `terminal-'.
454 It turns out that most of the offending Lisp functions were defined
455 in the trunk. Therefore, compatibility aliases should be defined
456 for the following names:
458 display-color-cells terminal-color-cells
459 display-color-p terminal-color-p
460 display-graphic-p terminal-graphic-p
461 display-grayscale-p terminal-grayscale-p
462 display-images-p terminal-images-p
463 display-mm-height terminal-mm-height
464 display-mm-width terminal-mm-width
465 display-mouse-p terminal-mouse-p
466 display-multi-font-p terminal-multi-font-p
467 display-multi-frame-p terminal-multi-frame-p
468 display-pixel-height terminal-pixel-height
469 display-pixel-width terminal-pixel-width
470 display-pixels-per-inch terminal-pixels-per-inch
471 display-planes terminal-planes
472 display-popup-menus-p terminal-popup-menus-p
473 display-save-under terminal-save-under
474 display-screens terminal-screens
475 display-supports-face-attributes-p terminal-supports-face-attributes-p
476 display-visual-class terminal-visual-class
477 framep-on-display framep-on-terminal
478 frames-on-display-list frames-on-terminal-list
480 The following functions were introduced in the multi-tty branch, and
481 can be renamed without aliases:
483 display-controlling-tty-p terminal-controlling-tty-p
484 display-list terminal-list
485 display-live-p terminal-live-p
486 display-name terminal-name
487 display-tty-type terminal-tty-type
488 frame-display terminal-of-frame
489 delete-display delete-terminal
491 ** The semantics of terminal-local variables are confusing; it is not
492 clear what binding is in effect in any given time. See if
493 current_kboard (or at least the terminal-local bindings exported to
494 Lisp) might be changed to be tied to the selected frame instead.
495 Currently, `function-key-map' and `key-translation-map' may be
496 accessed reliably only using the hackish
497 `(set-)terminal-local-value' functions.
499 Perhaps there should be a difference between `last-command' &co.
500 and these more conventional configuration variables.
501 (E.g. `symbol-value' would use current_kboard to access
502 `last-command', but SELECTED_FRAME()->display->kboard to get the
503 value of `function-key-map'.
505 ** The single-keyboard mode of MULTI_KBOARD is extremely confusing
506 sometimes; Emacs does not respond to stimuli from other keyboards.
507 At least a beep or a message would be important, if the single-mode
508 is still required to prevent interference. (Reported by Dan
511 Update: selecting a region with the mouse enables single_kboard
512 under X. This is very confusing.
514 Update: After discussions with Richard Stallman, this will be
515 resolved by having locked displays warn the user to wait, and
516 introducing a complex protocol to remotely bail out of
517 single-kboard mode by pressing C-g.
519 Update: Warning the user is not trivial to implement, as Emacs has
520 only one echo area, shared by all frames. Ideally the warning
521 should not be displayed on the display that is locking the others.
522 Perhaps the high probability of user confusion caused by
523 single_kboard mode deserves a special case in the display code.
524 Alternatively, it might be good enough to signal single_kboard mode
525 by changing the modelines or some other frame-local display element
526 on the locked out displays.
528 ** The session management module is prone to crashes when the X
529 connection is closed and then later I try to connect to a new X
532 #0 0xb7ebc806 in SmcGetIceConnection () from /usr/X11R6/lib/libSM.so.6
533 #1 0x080e6641 in x_session_check_input (bufp=0xbf86c9c0) at xsmfns.c:144
534 #2 0x080d3bbc in XTread_socket (device=0xa722ff8, expected=1, hold_quit=0xbf86ca90) at xterm.c:7037
535 #3 0x080fa404 in read_avail_input (expected=1) at keyboard.c:6696
536 #4 0x080fa4ca in handle_async_input () at keyboard.c:6900
537 #5 0x080d51fa in x_term_init (display_name=162628899, xrm_option=0x0, resource_name=0x857068c "emacs") at xterm.c:10622
538 #6 0x080d920e in x_display_info_for_name (name=162628899) at xfns.c:3975
539 #7 0x080d92f9 in check_x_display_info (object=1) at xfns.c:274
540 #8 0x080d97b8 in Fx_create_frame (parms=151221485) at xfns.c:3016
541 #9 0x0815bf72 in Ffuncall (nargs=2, args=0xbf86ceec) at eval.c:2851
543 I installed a workaround to prevent this. The X session manager is
544 only contacted when the very first display in the Emacs session is
545 an X display. Also, x_delete_display() on this display aborts
546 session management, and XTread_socket only calls
547 x_session_check_input when it is called for the display that the
548 session was opened on. While this does not really fix the bug, it
549 makes it much less frequent, because session manager support will
550 not normally be enabled when Emacs can survive the shutdown of the
553 See if xsmfns.c should be updated.
555 ** normal-erase-is-backspace-mode in simple.el needs to be updated for
556 multi-tty (rep. by Dan Waber).
558 ** Hunt down display-related functions in frame.el and extend them all
559 to accept display ids.
561 ** rif->flush_display_optional (NULL) calls should be replaced by a
564 ** The set-locale-environment hack (adding the DISPLAY option) should
565 be replaced with a clean design.
567 ** standard-display-table should be display-local.
568 standard-display-european should be display-local.
570 ** With iswitchb-default-method set to 'always-frame, only frames on
571 the current display should be considered. This might involve
572 extending `get-buffer-window'.
574 ** Have a look at Vlocale_coding_system. Seems like it would be a
575 tedious job to localize it, although most references use it for
576 interfacing with libc and are therefore OK with the global
579 Exceptions found so far: x-select-text and
580 x-cut-buffer-or-selection-value.
582 ** Have a look at fatal_error_hook.
584 ** Have a look at set_frame_matrix_frame.
586 ** Check if we got term-setup-hook right.
588 ** I think tip_frame should be display-local.
590 ** Check display reference count handling in x_create_tip_frame.
592 ** make-frame does not correctly handle extra parameters in its
595 (frame-parameter (make-frame (list (cons 'foobar 42))) 'foobar)
598 (This is likely an error in the CVS trunk.)
600 ** Dan Nicolaescu suggests that -nw should be added as an alias for -t
601 in emacsclient. Good idea. (Alas, implementing this is not
602 trivial, getopt_long does not seem to support two-letter ``short''
603 options. Patches are welcome.) :-)
605 ** Mark Plaksin suggests that emacsclient should accept the same
606 X-related command-line arguments as Emacs. Most of the X-related
607 argument-handling is done in Lisp, so this should be quite easy to
610 ** Gergely Nagy suggests that C-x # should only kill the current
611 frame, not any other emacsclient frame that may have the same file
612 opened for editing. I think I agree with him.
614 ** Very strange bug: visible-bell does not work on secondary
615 terminals in xterm and konsole. The screen does flicker a bit,
616 but it's so quick it isn't noticable.
618 ** Move baud_rate to struct display.
620 ** Implement support for starting an interactive Emacs session without
621 an initial frame. (The user would connect to it and open frames
622 later, with emacsclient.)
624 ** Fix Mac support (I can't do this entirely myself). Note that the
625 current state of Mac-specific source files in the multi-tty tree
626 are not useful; before starting work on Mac support, revert to
627 pristine, pre-multi-tty versions.
629 ** Fix W32 support (I can't do this entirely myself). Note that the
630 current state of W32-specific source files in the multi-tty tree
631 are not useful; before starting work on W32 support, revert to
632 pristine, pre-multi-tty versions.
634 ** Fix DOS support (I can't do this entirely myself). Note that the
635 current state of DOS-specific source files in the multi-tty tree
636 are not useful; before starting work on DOS support, revert to
637 pristine, pre-multi-tty versions.
639 ** Do a grep on XXX and ?? for more issues.
641 ** Understand Emacs's low-level input system (it's black magic) :-)
642 What exactly does interrupt_input do? I tried to disable it for
643 raw secondary tty support, but it does not seem to do anything
644 useful. (Update: Look again. X unconditionally enables this, maybe
645 that's why raw terminal support is broken again. I really do need
646 to understand input.)
648 ** Fix stuff_char for multi-tty. Doesn't seem to be of high priority.
653 (ex-TODO items with explanations.)
655 -- Introduce a new struct for terminal devices.
657 (Done, see struct tty_output. The list of members is not yet
660 -- Change the bootstrap procedure to initialize tty_list.
662 (Done, but needs review.)
664 -- Change make-terminal-frame to support specifying another tty.
666 (Done, new frame parameters: `tty' and `tty-type'.)
668 -- Implement support for reading from multiple terminals.
670 (Done, read_avail_input tries to read from each terminal, until one
671 succeeds. MULTI_KBOARD is not used. Secondary terminals don't send
674 (Update: They do, now.)
676 (Update2: After enabling X, they don't.)
678 -- other-frame should cycle through the frames on the `current'
681 (Done, by trivially modifiying next_frame and prev_frame.)
683 -- Support different terminal sizes.
687 -- Make sure terminal resizes are handled gracefully. (Could be
690 (Done. We don't get automatic SIGWINCH for additional ttys,
693 -- Extend emacsclient to automatically open a new tty when it connects
696 (Done. It's an ugly hack, needs more work.)
698 -- Redisplay must refresh the topmost frame on *all* terminals, not
699 just the initial terminal.
701 (Done, but introduced an ugly redisplay problems. Ugh.)
703 -- Fix redisplay problems.
705 (Done; it turned out that the entire Wcm structure must be moved
706 inside tty_output. Why didn't I catch this earlier?)
708 -- Provide a way for emacsclient to tell Emacs that the tty has been
711 (Done, simply forward the SIGWINCH signal.)
713 -- Each keypress should automatically select the frame corresponding
714 to the terminal that it was coming from. This means that Emacs
715 must know from which terminal the last keyboard event came from.
717 (Done, it was quite simple, the input event system already
718 supported multiple frames.)
720 -- Fix SIGIO issue with secondary terminals.
722 (Done, emacsclient signals Emacs after writing to the proxy pseudo
723 terminal. Note that this means that multi-tty does not work with
726 (Update: This is bullshit. There is a read_input_waiting function,
727 extend that somehow.)
729 (Update of update: The first update was not right either, extending
730 read_input_waiting was not necessary. Secondary ttys do seem to
731 send signals on input.)
733 (Update^3: Not any more.)
735 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
736 parameters from the currently selected terminal before the global
741 -- Put all cached terminal escape sequences into struct tty_output.
742 Currently, they are still stored in global variables, so we don't
743 really support multiple terminal types.
745 (Done. It was not fun.)
747 -- Implement sane error handling after initialization. (Currently
748 emacs exits if you specify a bad terminal type.) The helpful error
749 messages must still be provided when Emacs starts.
753 -- Implement terminal deletion, i.e., deleting local frames, closing
754 the tty device and restoring its previous state without exiting
757 (Done, but at the moment only called when an error happens during
758 initialization. There is a memory corruption error around this
759 somewhere.) (Update: now it is fully enabled.)
761 -- Implement automatic deletion of terminals when the last frame on
762 that terminal is closed.
766 -- Restore tty screen after closing the terminal.
768 (Done, we do the same as Emacs 21.2 for all terminals.)
770 -- 'TERM=dumb src/emacs' does not restore the terminal state.
774 -- C-g should work on secondary terminals.
776 (Done, but the binding is not configurable.)
778 -- Deal with SIGHUP in Emacs and in emacsclient. (After this, the
779 server-frames may be removed from server.el.)
781 (Done, nothing to do. It seems that Emacs does not receive SIGHUP
782 from secondary ttys, which is actually a good thing.) (Update: I
783 think it would be a bad idea to remove server-frames.)
785 -- Change emacsclient/server.el to support the -t argument better,
786 i.e. automatically close the socket when the frame is closed.
788 (Seems to be working OK.)
790 -- Fix mysterious memory corruption error with tty deletion. To
791 trigger it, try the following shell command:
793 while true; do TERM=no-such-terminal-definition emacsclient -h; done
795 Emacs usually dumps core after a few dozen iterations. (The bug
796 seems to be related to the xfreeing or bzeroing of
797 tty_output.Wcm. Maybe there are outside references to struct Wcm?
798 Why were these vars collected into a struct before multi-tty
801 (Done. Whew. It turned out that the problem had nothing to do
802 with hypothetical external references to Wcm, or any other
803 tty_output component; it was simply that delete_tty closed the
804 filehandles of secondary ttys twice, resulting in fclose doubly
805 freeing memory. Utterly trivial matter. I love the C's memory
806 management, it puts hair on your chest.)
808 -- Support raw secondary terminals. (Note that SIGIO works only on
809 the controlling terminal.) Hint: extend read_input_waiting for
810 multiple ttys and hopefully this will be fixed.
812 (Done, it seems to have been working already for some time. It
813 seems F_SETOWN does work, after all. Not sure what made it fail
814 earlier, but it seems to be fixed (there were several changes
815 around request_sigio, maybe one of them did it).
816 read_input_waiting is only used in sys_select, don't change
817 it.) (Update: After adding X support, it's broken again.)
818 (Update^2: No it isn't.) :-)
820 -- Find out why does Emacs abort when it wants to close its
821 controlling tty. Hint: chan_process[] array. Hey, maybe
822 noninterrupt-IO would work, too? Update: no, there is no process
825 (Done. Added add/delete_keyboard_wait_descriptor to
826 term_init/delete_tty. The hint was right, in a way.)
828 -- Issue with SIGIO: it needs to be disabled during redisplay. See if
829 fcntl kernel behaviour could be emulated by emacsclient.
831 (Done. Simply disabled the SIGIO emulation hack in emacsclient.)
832 (Update: it was added back.) (Update^2: and removed again.)
834 -- server.el: There are issues with saving files in buffers of closed
835 clients. Try editing a file with emacsclient -f, and (without
836 saving it) do a delete-frame. The frame is closed without
837 question, and a surprising confirmation prompt appears in another
840 (Done. delete-frame now asks for confirmation if it still has
841 pending buffers, and modified buffers don't seem to be deleted.)
843 -- emacsclient.el, server.el: Handle eval or file open errors when
848 -- Make parts of struct tty_output accessible from Lisp. The device
849 name and the type is sufficient.
851 (Done, see frame-tty-name and frame-tty-type.)
853 -- Export delete_tty to the Lisp environment, for emacsclient.
855 (Done, see delete-tty.)
857 -- Get rid of the accessor macros in termchar.h, or define macros for
862 -- Move device-specific parameters (like costs) commonly used by
863 device backends to a common, device-dependent structure.
865 (Done. See struct display_method in termhooks.h.)
869 (Done. Well, it seems to be working.)
871 -- Allow simultaneous X and tty frames. (Handling input could be
872 tricky. Or maybe not.)
874 (Done. Allowed, that is. It is currently extremely unstable, to
875 the point of being unusable. The rif variable causes constant
876 core dumps. Handling input is indeed tricky.)
878 -- Rewrite multi-tty input in terms of MULTI_KBOARD.
880 (Done. In fact, there was no need to rewrite anything, I just
881 added a kboard member to tty_display_info, and initialized the
882 frame's kboard from there.)
884 -- Fix rif issue with X-tty combo sessions. IMHO the best thing to do
885 is to get rid of that global variable (and use the value value in
886 display_method, which is guaranteed to be correct).
888 (Done, did exactly that. Core dumps during combo sessions became
889 much rarer. In fact, I have not yet met a single one.)
891 -- Add multi-tty support to talk.el.
895 -- Clean up the source of emacsclient. It is a mess.
897 (Done, eliminated stupid proxy-pty kludge.)
899 -- Fix faces on tty frames during X-tty combo sessions. There is an
900 init_frame_faces call in init_sys_modes, see if there is a problem
903 (Done, there was a stupid mistake in
904 Ftty_supports_face_attributes_p. Colors are broken, though.)
906 -- C-x 5 2, C-x 5 o, C-x 5 0 on an emacsclient frame unexpectedly
907 exits emacsclient. This is a result of trying to be clever with
908 delete-frame-functions.
910 (Fixed, added delete-tty-after-functions, and changed server.el to
913 -- Something with (maybe) multi-keyboard support broke function keys
914 and arrows on ttys during X+tty combo sessions. Debug this.
916 (I can't reproduce it, maybe the terminal type was wrong.)
918 -- Fix input from raw ttys (again).
920 (Now it seems to work all right.)
922 -- During an X-tty combo session, a (message "Hello") from a tty frame
923 goes to the X frame. Fix this.
925 (Done. There was a safeguard against writing to the initial
926 terminal frame during bootstrap which prevented echo_area_display
927 from working correctly on a tty frame during a combo session.)
929 -- If there are no frames on its controlling terminal, Emacs should
930 exit if the user presses C-c there.
932 (Done, as far as possible. See the SIGTERM comment in
933 interrupt_signal on why this seems to be impossible to solve this
936 -- During an X session, Emacs seems to read from stdin. Also, Emacs
937 fails to start without a controlling tty.
939 (Fixed by replacing the troublesome termcap display with a dummy
940 bootstrap display during bootstrap.
942 -- Do tty output through struct display, like graphical display
947 -- Define an output_initial value for output_method for the initial
948 frame that is dumped with Emacs. Checking for this frame (e.g. in
949 cmd_error_internal) is ugly.
951 (Done, broking interactive temacs.)
953 -- The command `emacsclient -t -e '(delete-frame)'' fails to exit.
957 -- frame-creation-function should always create a frame that is on the
958 same display as the selected frame. Maybe frame-creation-function
959 should simply be removed and make-frame changed to do the right
962 (Done, with a nice hack. frame-creation-function is now frame-local.)
964 -- Fix C-g on raw ttys.
966 (Done. I disabled the interrupt/quit keys on all secondary
967 terminals, so Emacs sees C-g as normal input. This looks like an
968 overkill, because emacsclient has extra code to pass SIGINT to
969 Emacs, so C-g should remain the interrupt/quit key on emacsclient
970 frames. See the next entry why implementing this distinction would
973 -- Make sure C-g goes to the right frame with ttys. This is hard, as
974 SIGINT doesn't have a tty parameter. :-(
976 (Done, the previous change fixes this as a pleasant side effect.)
978 -- I have seen a case when Emacs with multiple ttys fell into a loop
979 eating 100% of CPU time. Strace showed this loop:
982 kill(30284, SIGIO) = 0
983 --- SIGIO (I/O possible) @ 0 (0) ---
984 ioctl(6, FIONREAD, [0]) = -1 EIO (Input/output error)
985 ioctl(5, FIONREAD, [0]) = -1 EIO (Input/output error)
986 ioctl(0, FIONREAD, [0]) = 0
987 sigreturn() = ? (mask now [])
988 gettimeofday({1072842297, 747760}, NULL) = 0
989 gettimeofday({1072842297, 747806}, NULL) = 0
990 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
991 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
992 gettimeofday({1072842297, 748245}, NULL) = 0
994 I have seen something similar with a single X frame, but have not
995 been able to reproduce it for debugging.
997 Update: This may have been caused by checking for nread != 0
998 instead of nread > 0 after calling read_socket_hook in
1001 (Fixed. This was caused by unconditionally including stdin in
1002 input_wait_mask in init_process. The select call in
1003 wait_reading_process_input always returned immediately, indicating
1004 that there is pending input from stdin, which nobody read.
1006 Note that the above strace output seems to be an unrelated but
1007 similar bug. I think that is now fixed.)
1009 -- Exiting Emacs while there are emacsclient frames doesn't restore the
1010 ttys to their default states.
1012 (This seems to be fixed by some previous change.)
1014 -- Allow opening an X session after -nw.
1018 -- Fix color handling during tty+X combo sessions. (It seems that tty
1019 sessions automatically convert the face colors to terminal colors
1020 when the face is loaded. This conversion must happen instead on
1021 the fly in write_glyphs, which might be problematic, as color
1022 approximation is currently done in lisp (term/tty-colors.el).)
1023 (Update: hm, colors seem to work fine if I start emacs with -nw and
1024 then create an X frame. Maybe it's just a small buglet somewhere.)
1026 (Seems to be fixed. The problem was in startup.el, it did not
1027 initialize tty colors when the initial window system was
1030 -- emacs -nw --eval '(y-or-n-p "Foobar")' segfaults. (Reported by
1033 (Fixed, there was a keyboard initialization problem.)
1035 -- Fix interactive use of temacs. There are face-related SEGVs, most
1036 likely because of changes in realize_default_face, realize_face.
1040 -- Don't exit Emacs when the last X connection fails during a
1041 multi-display session.
1045 -- Dan Nicolaescu noticed that starting emacsclient on the same
1046 terminal device that is the controlling tty of the Emacs process
1047 gives unexpected results.
1051 -- Istvan Marko reported that Emacs hang on ttys if it was started
1052 from a shell script.
1054 (Fixed. There was a bug in the multi-tty version of
1055 narrow_foreground_group. tcsetpgrp blocks if it is called from a
1056 process that is not in the same process group as the tty.)
1058 -- emacsclient -t from an Emacs term buffer does not work, complains
1059 about face problems. This can even lock up Emacs (if the recursive
1060 frame sets single_kboard). Update: the face problems are caused by
1061 bugs in term.el, not in multi-tty. The lockup is caused by
1062 single_kboard mode, and is not easily resolvable. The best thing to
1063 do is to simply refuse to create a tty frame of type `eterm'.
1065 (Fixed, changed emacsclient to check for TERM=eterm. The face
1066 complaints seem to be caused by bugs in term.el; they are not
1067 related to multi-tty.)
1069 -- Find out the best way to support suspending Emacs with multiple
1070 ttys. My guess: disable it on the controlling tty, but from other
1071 ttys pass it on to emacsclient somehow. (It is (I hope) trivial to
1072 extend emacsclient to handle suspend/resume. A `kill -STOP' almost
1075 (Done. I needed to play with signal handling and the server
1076 protocol a bit to make emacsclient behave as a normal UNIX program
1077 wrt foreground/background process groups.)
1079 -- There is a flicker during the startup of `emacs -nw'; it's as if
1080 the terminal is initialized, reset and then initialialized again.
1081 Debug this. (Hint: narrow_foreground_group is called twice during
1086 -- Robert Chassell has found serious copy-paste bugs with the
1087 multi-tty branch. There seem to be redisplay bugs while copying
1088 from X to a terminal frame. Copying accented characters do not
1091 (Patch-124 should fix this, by changing the interprogram-*-function
1092 variables to be frame-local, as suggested by Mark Plaksin
1093 (thanks!). I think that the redisplay bugs are in fact not bugs,
1094 but delays caused by single_kboard --> perhaps MULTI_KBOARD should
1097 -- frame-creation-function was removed, which might be a bad idea.
1098 Think up a compatible solution.
1100 (It was an internal interface that may be changed when necessary.)
1102 -- Change Lisp code not to (getenv "TERM"); use the `tty-type' frame
1103 parameter or the frame-tty-type function instead. (M-x tags-search
1104 "TERM" helps with this.) Update: Actually, all getenv invocations
1105 should be checked for multi-tty compatibility, and an interface
1106 must be implemented to get the remote client's environment.
1108 (Done. Only getenv calls in lisp/term/*.el were changed; other
1109 calls should be mostly left as they are.)
1111 -- Add an elaborate mechanism for display-local variables. (There are
1112 already a few of these; search for `terminal-local' in the Elisp
1115 (Not needed. Display-local variables could be emulated by
1116 frame-local variables.)
1118 -- Emacs assumes that all terminal frames have the same locale
1119 settings as Emacs itself. This may lead to bogus results in a
1120 multi-locale setup. (E.g., while logging in from a remote client
1121 with a different locale.)
1122 (Update after new bugreport by Friedrich Delgado Friedrichs:
1123 (at least) the structs terminal_coding and keyboard_coding in
1124 coding.c must be moved to struct display, and the Lisp interface
1125 [set-]keyboard-coding-system must be adapted for the change.)
1127 (Fixed. Emacs now uses the locale settings as seen by the
1128 emacsclient process for server tty frames.)
1129 (Update: Not really; Vlocale_coding_system is still global.)
1131 -- Make `struct display' accessible to Lisp programs. Accessor functions:
1133 (displayp OBJECT): Returns t if OBJECT is a display.
1134 => Implemented as display-live-p.
1136 (display-list): Returns list of currently active displays.
1139 (selected-display): Returns the display object of the selected frame.
1140 => Not strictly necessary, but implemented anyway.
1142 (frame-display FRAME): Returns the display object of FRAME.
1145 (display-frames DISPLAY): Returns a list of frames on DISPLAY.
1146 => Already implemented, see frames-on-display-list.
1148 (display-type DISPLAY): Returns the type of DISPLAY, as a
1149 symbol. (See `framep'.)
1150 => Implemented as display-live-p.
1152 (display-device DISPLAY): Returns the name of the device that
1153 DISPLAY uses, as a string. (E.g: "/dev/pts/16", or
1155 => Implemented as display-name.
1159 See next issue why this is necessary.
1161 (Update: The consensus on emacs-devel seems to be to do this via
1162 integer identifiers. That's fine by me.)
1166 -- The following needs to be supported:
1173 The cleanest way to solve this is to allow multiple displays on the
1174 same terminal device; each new emacsclient process should create
1175 its own display. As displays are currently identified by their
1176 device names, this is not possible until struct display becomes
1177 accessible as a Lisp-level object.
1181 -- Miles Bader suggests that C-x C-c on an emacsclient frame should
1182 only close the frame, not exit the entire Emacs session. Update:
1183 see above for a function that does this. Maybe this should be the
1186 (Done. This is the new default. No complaints so far.)
1188 -- Clean up the frame-local variable system. I think it's ugly and
1189 error-prone. But maybe I just haven't yet fully understood it.
1191 (Nothing to do. It doesn't seem ugly any more. It's rather clever.)
1193 -- Support multiple character locales. A version of
1194 `set-locale-environment' needs to be written for setting up
1195 display-local settings on ttys. I think calling
1196 set-display-table-and-terminal-coding-system and
1197 set-keyboard-coding-system would be enough. The language
1198 environment itself should remain a global setting.
1200 (Done, by an ugly hack.)
1202 -- The terminal customization files in term/*.el tend to change global
1203 parameters, which may confuse Emacs with multiple displays. Change
1204 them to tweak only frame-local settings, if possible. (They tend
1205 to call define-key to set function key sequences a lot.)
1207 (Done, by making `function-key-map' terminal-local (i.e., part of
1208 struct kboard). This has probably covered all the remaining problems.)
1210 -- Make `function-key-map' and `key-translation-map' terminal-local.
1214 -- Implement `terminal-local-value' and `set-terminal-local-value' to
1215 allow deterministic access to terminal local bindings. The
1216 encode-kb package can not set up `key-translation-map' without
1217 these. The terminal-local bindings seem to be independent of what
1222 -- xt-mouse.el needs to be adapted for multi-tty. It currently
1223 signals an error on kill-emacs under X, which prevents the user
1224 from exiting Emacs. (Reported by Mnemonikk on freenode.)
1229 ;;; arch-tag: 8da1619e-2e79-41a8-9ac9-a0485daad17d