Fix permissions. (Reported by Han Boetes.)
[bpt/emacs.git] / README.multi-tty
1 -*- coding: utf-8; mode: text; -*-
2 GOAL
3 ----
4
5 This branch implements support for opening multiple, different tty
6 devices and simultaneous X and tty frames from a single Emacs session.
7
8 Some use cases:
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 tty.
13 Now, with multi-tty support, it can do that. (Emacsclient starts up
14 faster than vi!)
15
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
20 to gnus-slave.
21
22 WHO IS DOING IT
23 ---------------
24
25 I'm Károly Lőrentey. My address: lorentey@elte.hu.
26
27 Comments, bug reports, suggestions and patches are welcome; send them
28 to multi-tty@lists.fnord.hu.
29
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!
33
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 mace <mace@kirjakaapeli.lib.hel.fi>
50 Xavier Mallard <zedek@gnu-rox.org>
51 Istvan Marko <mi-mtty@kismala.com>
52 Ted Morse <morse@ciholas.com>
53 Gergely Nagy <algernon@debian.org>
54 Dan Nicolaescu <dann@ics.uci.edu>
55 Kalle Olavi Niemitalo <kon@iki.fi>
56 Mark Plaksin <happy@mcplaksin.org>
57 Frank Ruell <stoerte@dreamwarrior.net>
58 Tom Schutzer-Weissmann <trmsw@yahoo.co.uk>
59 Joakim Verona <joakim@verona.se>
60 Dan Waber <dwaber@logolalia.com>
61 and many others.
62
63 Richard Stallman was kind enough to review an earlier version of my
64 patches.
65
66
67 MAILING LISTS
68 -------------
69
70 The multi-tty mailing list (discussion & bug reports):
71
72 Address: multi-tty@lists.fnord.hu
73 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty
74 Archive: http://news.gmane.org/gmane.emacs.multi-tty/
75
76 Commit notifications (read-only):
77
78 Address: multi-tty-commits@lists.fnord.hu
79 Signup: http://lists.fnord.hu/mailman/listinfo/multi-tty-commits
80
81
82 STATUS
83 ------
84
85 The branch is now very stable and almost full-featured. All of the
86 major problems have been fixed, only a few minor issues remain. (It
87 still needs to be ported to Windows/Mac/DOS, though.) Both multiple
88 tty device support and simultaneous X and tty frame support works
89 fine. Emacsclient has been extended to support opening new tty and X
90 frames. It has been changed to open new Emacs frames by default.
91
92 The multi-tty branch has been scheduled for inclusion in the next
93 major release of Emacs (version 23). I expect the merge into the
94 development trunk to occur sometime during next year (2006), after the
95 merge of the Unicode branch.
96
97 Tested on GNU/Linux, Solaris 8, FreeBSD and OpenBSD. Please let me
98 know if you succeed or fail to use it on other platforms---I'll have a
99 few tricky test cases for you.
100
101 Known problems:
102
103 * GTK support. If you compile your Emacs with the GTK
104 toolkit, some functionality of multi-tty will be lost. In
105 particular, you will not be able to work on multiple X
106 displays at once. Current releases of GTK have limitations
107 and bugs that prevent full-blown multi-display support in
108 Emacs. (GTK crashes when Emacs tries to disconnect from an
109 X server.) Use the Lucid toolkit if you want to see a
110 complete feature set.
111
112 * The single-kboard mode.
113
114 If your multi-tty Emacs session seems to be frozen, you
115 probably have a recursive editing session or a pending
116 minibuffer prompt (which is a kind of recursive editing) on
117 another display. To unfreeze your session, switch to that
118 display and complete the recursive edit, for example by
119 pressing C-] (`abort-recursive-edit').
120
121 I am sorry to say that currently there is no way to break
122 out of this "single-kboard mode" from a frozen display. If
123 you are unable to switch to the display that locks the
124 others (for example because it is on a remote computer),
125 then you can use emacsclient to break out of all recursive
126 editing sessions:
127
128 emacsclient -e '(top-level)'
129
130 Note that this (perhaps) unintuitive behaviour is by design.
131 Single-kboard mode is required because of an intrinsic Emacs
132 limitation that is very hard to eliminate. (This limitation
133 is related to the single-threaded nature of Emacs.)
134
135 I plan to implement better user notification and support for
136 breaking out of single-kboard mode from locked displays.
137
138 * Mac, Windows and DOS support is broken, doesn't even
139 compile. Multiple display support will probably not provide
140 new Emacs features on these systems, but the multi-tty
141 branch changed a few low-level interfaces, and the
142 system-dependent source files need to be adapted
143 accordingly. The changes are mostly trivial, so almost
144 anyone can help, if only by compiling the branch and
145 reporting the compiler errors.
146
147
148 HOW TO GET THE BRANCH
149 ---------------------
150
151 The branch uses Bazaar 1 (http://bazaar.canonical.com) for version control.
152
153 Retrieving the latest version of the branch:
154
155 baz register-archive -f http://aszt.inf.elte.hu/~lorentey/mirror/arch/2004
156 baz get lorentey@elte.hu--2004/emacs--multi-tty <directory>
157
158 This incantation uses an archive mirror that is hosted on a
159 high-bandwidth site. Please note that on average there is a two-hour
160 delay for commits to arrive on this mirror. My primary mirror is on the
161 low-bandwidth http://lorentey.hu/ site:
162
163 baz register-archive -f http://lorentey.hu/arch/2004/
164 baz get lorentey@elte.hu--2004/emacs--multi-tty <directory>
165
166 This is "instantly" updated, but very slow from outside Hungary.
167 (By "instantly" I mean as soon as I connect the notebook I work on to
168 a network. It could take days.)
169
170 The Arch supermirror provides mirroring services for all public Arch
171 repositories. We have a mirror there, too, if you prefer.
172
173 baz register-archive -f http://mirrors.sourcecontrol.net/lorentey%40elte.hu--2004
174 baz get lorentey@elte.hu--2004/emacs--multi-tty <directory>
175
176 My GPG key id is 0FB27A3F; it is available from
177 hkp://wwwkeys.eu.pgp.net/, or from my homepage at
178 http://lorentey.hu/rolam/gpg.html)
179
180 Don't worry if the above checkout takes a few minutes to complete;
181 once you have a source tree, updating it to the latest revision will
182 be _much_ faster. Use the following command for the update:
183
184 baz replay
185
186 You can find more information about Bazaar on
187 http://bazaar.canonical.com/. It's a distributed source control
188 system that is somewhat less broken than competing projects.
189
190 If you don't have Bazaar, the branch has a homepage from which you can
191 download conventional patches against Emacs CVS HEAD:
192
193 http://lorentey.hu/project/emacs.html
194
195 I suggest you use Bazaar whenever feasible.
196
197
198 DEBIAN PACKAGES
199 ---------------
200
201 If you run Debian, or a distribution based on Debian, you are welcome
202 to use our binary packages; put these lines in your /etc/apt/sources.list:
203
204 # Multi-tty Emacs
205 deb http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
206 deb-src http://aszt.inf.elte.hu/~lorentey/mirror/apt unstable multi-tty
207
208 Note that these packages are intended solely to provide an easy way to
209 test the new multi-tty features. They are not to be taken as Emacs
210 releases, and it's a mistake to expect robust operation or any kind of
211 timely support for them. Do not install them, unless you'd like to
212 have your editor crash on you.
213
214
215 COMPILATION
216 -----------
217
218 The multi-tty branch is compiled the same way as Emacs itself:
219
220 make maintainer-clean # (If you have compiled Emacs before)
221
222 ./configure --without-gtk <your favourite options>
223 make bootstrap
224 make install
225
226 If you have strange compilation errors, they may be caused by old
227 *.elc files that are left over from an earlier bootstrap. The `make
228 maintainer-clean' target deletes them, so it is a good idea to run
229 that before reporting a bug. (Emacs requires a clean recompilation
230 after certain kinds of source code changes.)
231
232 TESTING
233 -------
234
235 To test the multi-tty branch, start up the Emacs server with the
236 following commands:
237
238 emacs
239 M-x server-start
240
241 and then (from a shell prompt on another terminal) start emacsclient
242 with
243 emacsclient -t /optional/file/names... (for a tty frame)
244 emacsclient /optional/file/names... (for an X frame)
245
246 (Make sure both emacs and emacsclient are multi-tty versions.)
247 You'll hopefully have two fully working, independent frames on
248 separate terminals. The new frame is closed automatically when you
249 finish editing the specified files (C-x #), but delete-frame (C-x 5 0)
250 also works. Of course, you can create frames on more than two tty
251 devices.
252
253 Creating new frames on the same tty with C-x 5 2 (make-frame-command)
254 works, and behaves the same way as in previous Emacs versions. If you
255 exit emacs, all terminals should be restored to their previous states.
256
257 This is work in progress, and probably full of bugs. It is a good
258 idea to run emacs from gdb, so that you'll have a live instance to
259 debug if something goes wrong. Please send me your bug reports on our
260 mailing list: multi-tty@lists.fnord.hu
261
262 TIPS & TRICKS
263 -------------
264
265 I think the best way to use the new Emacs is to have it running inside
266 a disconnected GNU screen session, and always use emacsclient for
267 normal work. One advantage of this is that not a single keystroke of
268 your work will be lost if the display device that you are using
269 crashes, or the network connection times out, or whatever. (I had an
270 extremely unstable X server for some time while I was developing these
271 patches, and running Emacs this way has saved me a number of M-x
272 recover-session invocations.)
273
274 I use the following two bash scripts to handle my Emacs sessions:
275
276 -------------------------------------------------------connect-emacs--
277 #!/bin/bash
278 # Usage: connect-emacs <name> <args>...
279 #
280 # Connects to the Emacs instance called NAME. Starts up the instance
281 # if it is not already running. The rest of the arguments are passed
282 # to emacsclient.
283
284 name="$1"
285 shift
286
287 if [ -z "$name" ]; then
288 echo "Usage: connect_emacs <name> <args>..." >&2
289 exit 1
290 fi
291 preload-emacs "$name" wait
292 /usr/bin/emacsclient.emacs-multi-tty -s "$name" "$@"
293 ----------------------------------------------------------------------
294
295 -------------------------------------------------------preload-emacs--
296 #!/bin/bash
297 # Usage: preload-emacs <name> [<waitp>]
298 #
299 # Preloads the Emacs instance called NAME in a detached screen
300 # session. Does nothing if the instance is already running. If WAITP
301 # is non-empty, the function waits until the server starts up and
302 # creates its socket; otherwise it returns immediately.
303
304 name="$1"
305 waitp="$2"
306 screendir="/var/run/screen/S-$USER"
307 serverdir="/tmp/emacs$UID"
308 emacs=/usr/bin/emacs-multi-tty # Or wherever you installed your multi-tty Emacs
309
310 if [ -z "$name" ]; then
311 echo "Usage: preload_emacs <name> [<waitp>]" >&2
312 exit 1
313 fi
314
315 if [ ! -e "$screendir"/*."$name" ]; then
316 if [ -e "$serverdir/$name" ]; then
317 # Delete leftover socket (for the wait option)
318 rm "$serverdir/$name"
319 fi
320 screen -dmS "$name" "$emacs" -nw --eval "(setq server-name \"$name\")" -f server-start
321 fi
322 if [ ! -z "$waitp" ]; then
323 while [ ! -e "$serverdir/$name" ]; do sleep 0.1; done
324 fi
325 ----------------------------------------------------------------------
326
327 I have the following in my profile to have two instances automatically
328 preloaded for editing and email:
329
330 preload-emacs editor
331 preload-emacs gnus
332
333 It is useful to set up short aliases for connect-emacs. I use the
334 following:
335
336 alias edit="connect-emacs editor"
337 alias e=edit
338 alias et="connect-emacs editor -t"
339 alias gnus="connect-emacs gnus"
340
341
342 CHANGELOG
343 ---------
344
345 See arch logs.
346
347
348 NEWS
349 ----
350
351 For the NEWS file: (Needs much, much work)
352
353 ** Support for multiple terminal devices and simultaneous graphical
354 and tty frames has been added. You can test for the presence of
355 this feature in your Lisp code by testing for the `multi-tty'
356 feature.
357
358 *** The `window-system' variable has been made frame-local. The new
359 `initial-window-system' variable contains the `window-system'
360 value for the first frame.
361
362 *** You can specify a terminal device (`tty' parameter) and a terminal
363 type (`tty-type' parameter) to `make-terminal-frame'.
364
365 *** The new function `make-frame-on-tty' allows you to create a new
366 frame on another tty device interactively.
367
368 *** The function `make-frame-on-display' now works during a tty
369 session, and `make-frame-on-tty' works during a graphical session.
370
371 *** Emacsclient has been extended to support opening a new terminal
372 frame. Its behaviour has been changed to open a new Emacs frame by
373 default. Use the -c option to get the old behavior of opening
374 files in the currently selected Emacs frame.
375
376 *** C-z now invokes `suspend-frame', C-x C-c now invokes
377 `save-buffers-kill-terminal'.
378
379 *** New functions: frame-tty-name, frame-tty-type, delete-tty,
380 suspend-tty, resume-tty, terminal-id, terminal-parameters,
381 terminal-parameter, set-terminal-parameter,
382 modify-terminal-parameters, environment, let-environment
383
384 *** New variables: local-key-translation-map, local-function-key-map
385
386 *** The `keyboard-translate-table' variable and the terminal and
387 keyboard coding systems have been made terminal-local.
388
389 *** In addition to the global key-translation-map and
390 function-key-map, Emacs has terminal-local
391 local-key-translation-map and local-function-key-map variables,
392 and uses them instead of the global keymaps to set up translations
393 and function key sequences relevant to a specific terminal device.
394
395 *** talk.el has been extended for multiple tty support.
396
397 * * *
398
399 (The rest of this file consists of my development notes and as such it
400 is probably not very interesting for anyone else.)
401
402 THINGS TO DO
403 ------------
404
405 ** emacsclient -t on the console does not work after sudo:
406
407 # sudo lorentey
408 $ emacsclient -t
409 *ERROR*: Could not open file: /dev/tty1
410
411 The tty can be opened as /dev/tty by emacsclient, but not by Emacs.
412 This seems to be a serious problem. Currently my only idea is to
413 bring back the ugly pty proxy hack from the initial versions of
414 multi-tty. Suggestions would be appreciated.
415
416 ** Understand how `quit_throw_to_read_char' works, and fix any bugs
417 that come to light.
418
419 ** See if getcjmp can be eliminated somehow. Why does Emacs allow
420 asynchronous input processing while it's reading input anyway?
421
422 ** `delete-frame' events are handled by `special-event-map'
423 immediately when read by `read_char'. This is fine but it prevents
424 higher-level keymaps from binding that event to get notified of the
425 deleted frame.
426
427 Sometimes it would be useful for Lisp code to be notified of frame
428 deletions after they have happened, usually because they want to
429 clean up after the deleted frame. Not all frame-local states can
430 be stored as a frame parameter. For example,
431 `display-splash-screen' uses `recursive-edit' with a special keymap
432 override to create its buffer---and it leads to all kinds of
433 nastiness if Emacs stays in this recursive edit mode after the
434 frame containing the splash screen is deleted. Basically, the
435 splash-screen implementation wants to throw out of the recursive
436 edit when the frame is deleted; however, it is not legal to throw
437 from `delete-frame-functions' because `delete-frame' must not fail.
438 (Introducing `delete-frame-after-functions' would not help either
439 because `delete-frame' may not fail at that time either.)
440
441 Currently `fancy-splash-screens' installs a
442 `delete-frame-functions' hook that sets up a timer to exit the
443 recursive edit. This is an adequate solution, but it would perhaps
444 be better to have something like a `frame-deleted' event that could
445 be bound in the normal way.
446
447 ** Trouble: `setenv' doesn't actually set environment variables in the
448 Emacs process. This defeats the purpose of the elaborate
449 `server-with-environment' magic around the `tgetent' call in
450 `init_tty'. D'oh.
451
452 ** (Possibly) create hooks in struct device for creating frames on a
453 specific terminal, and eliminate the hackish terminal-related frame
454 parameters (display, tty, tty-type).
455
456 make_terminal_frame
457 create_tty_output
458
459 ** Decide whether to keep the C implementation of terminal parameters,
460 or revert to the previous, purely Lisp code. It turned out that
461 local environments do not need terminal parameters after all.
462
463 ** Move Fsend_string_to_terminal to term.c, and declare get_named_tty
464 as static, removing it from dispextern.h.
465 Move fatal to emacs.c and declare it somewhere.
466
467 ** Search for `suspend-emacs' references and replace them with
468 `suspend-frame', if necessary. Ditto for `save-buffers-kill-emacs'
469 vs. `save-buffers-kill-display'.
470
471 ** Emacs crashes when a tty frame is resized so that there is no space
472 for all its windows. (Tom Schutzer-Weissmann)
473
474 ** Report GTK multi-display problems to GTK maintainers. For extra
475 credit, fix them.
476
477 Currently you can not connect to new X displays when you compile
478 Emacs with GTK support. If you want to play around with GTK
479 multidisplay (and don't mind core dumps), you can edit src/config.h
480 and define HAVE_GTK_MULTIDISPLAY there by hand.
481
482 Update: Han reports that GTK+ version 2.8.9 almost gets display
483 disconnects right. GTK will probably be fully fixed by the time
484 multi-tty gets into the trunk.
485
486 ** Audit `face-valid-attribute-values' usage in customize and
487 elsewhere. Its return value depends on the current window system.
488 Replace static initializers using it with runtime functions. For
489 example, custom's buttons are broken on non-initial device types.
490
491 ** Possibly turn off the double C-g feature when there is an X frame.
492 C.f. (emacs)Emergency Escape.
493
494 ** frames-on-display-list should also accept frames.
495
496 ** Consider the `tty-type' frame parameter and the `display-tty-type'
497 function. They serve the exact same purpose. I think it may be
498 a good idea to eliminate one of them, preferably `tty-type'.
499
500 ** The handling of lisp/term/*.el, and frame creation in general, is a
501 big, big mess. How come the terminal-specific file is loaded by
502 tty-create-frame-with-faces? I don't think it is necessary to load
503 these files for each frame; once per terminal should be enough.
504 Update: lisp/term/*.el is not loaded repeatedly anymore, but
505 faces.el still needs to be cleaned up.
506
507 ** Fix frame-set-background-mode in this branch. It was recently
508 changed in CVS, and frame.el in multi-tty has not yet been adapted
509 for the changes. (It needs to look at
510 default-frame-background-mode.) (Update: maybe it is fixed now;
511 needs testing.) (Note that the byte compiler has this to say about
512 term/rxvt.el:)
513
514 term/rxvt.el:309:17:Warning: assignment to free variable
515 `default-frame-background-mode'
516
517 ** I think `(set-)terminal-local-value' and the terminal parameter
518 mechanism should be integrated into a single framework.
519
520 (Update: `(set-)terminal-local-value' is now eliminated, but the
521 terminal-local variables should still be accessible as terminal
522 parameters. This also applies to `display-name' and similar
523 functions.)
524
525 ** Add the following hooks: after-delete-frame-hook (for server.el,
526 instead of delete-frame-functions),
527 after-delete-terminal-functions, after-create-terminal-functions.
528
529 ** BULK RENAME: The `display-' prefix of new Lisp-level functions
530 conflicts with stuff like `display-time-mode'. Use `device-'
531 or `terminal-' instead. I think I prefer `terminal-'.
532
533 It turns out that most of the offending Lisp functions were defined
534 in the trunk. Therefore, compatibility aliases should be defined
535 for the following names:
536
537 display-color-cells terminal-color-cells
538 display-color-p terminal-color-p
539 display-graphic-p terminal-graphic-p
540 display-grayscale-p terminal-grayscale-p
541 display-images-p terminal-images-p
542 display-mm-height terminal-mm-height
543 display-mm-width terminal-mm-width
544 display-mouse-p terminal-mouse-p
545 display-multi-font-p terminal-multi-font-p
546 display-multi-frame-p terminal-multi-frame-p
547 display-pixel-height terminal-pixel-height
548 display-pixel-width terminal-pixel-width
549 display-pixels-per-inch terminal-pixels-per-inch
550 display-planes terminal-planes
551 display-popup-menus-p terminal-popup-menus-p
552 display-save-under terminal-save-under
553 display-screens terminal-screens
554 display-supports-face-attributes-p terminal-supports-face-attributes-p
555 display-visual-class terminal-visual-class
556 framep-on-display framep-on-terminal
557 frames-on-display-list frames-on-terminal-list
558
559 The following functions were introduced in the multi-tty branch, and
560 were renamed without aliases:
561
562 delete-display delete-terminal
563 display-controlling-tty-p controlling-tty-p
564 display-list terminal-list
565 display-live-p terminal-live-p
566 display-name terminal-name
567 display-tty-type tty-type
568 frame-display frame-terminal
569 selected-display selected-terminal
570
571 ** The single-keyboard mode of MULTI_KBOARD is extremely confusing
572 sometimes; Emacs does not respond to stimuli from other keyboards.
573 At least a beep or a message would be important, if the single-mode
574 is still required to prevent interference. (Reported by Dan
575 Nicolaescu.)
576
577 Update: selecting a region with the mouse enables single_kboard
578 under X. This is very confusing.
579
580 Update: After discussions with Richard Stallman, this will be
581 resolved by having locked displays warn the user to wait, and
582 introducing a complex protocol to remotely bail out of
583 single-kboard mode by pressing C-g.
584
585 Update: Warning the user is not trivial to implement, as Emacs has
586 only one echo area, shared by all frames. Ideally the warning
587 should not be displayed on the display that is locking the others.
588 Perhaps the high probability of user confusion caused by
589 single_kboard mode deserves a special case in the display code.
590 Alternatively, it might be good enough to signal single_kboard mode
591 by changing the modelines or some other frame-local display element
592 on the locked out displays.
593
594 Update: In fact struct kboard does have an echo_string slot.
595
596 ** The session management module is prone to crashes when the X
597 connection is closed and then later I try to connect to a new X
598 session:
599
600 #0 0xb7ebc806 in SmcGetIceConnection () from /usr/X11R6/lib/libSM.so.6
601 #1 0x080e6641 in x_session_check_input (bufp=0xbf86c9c0) at xsmfns.c:144
602 #2 0x080d3bbc in XTread_socket (device=0xa722ff8, expected=1, hold_quit=0xbf86ca90) at xterm.c:7037
603 #3 0x080fa404 in read_avail_input (expected=1) at keyboard.c:6696
604 #4 0x080fa4ca in handle_async_input () at keyboard.c:6900
605 #5 0x080d51fa in x_term_init (display_name=162628899, xrm_option=0x0, resource_name=0x857068c "emacs") at xterm.c:10622
606 #6 0x080d920e in x_display_info_for_name (name=162628899) at xfns.c:3975
607 #7 0x080d92f9 in check_x_display_info (object=1) at xfns.c:274
608 #8 0x080d97b8 in Fx_create_frame (parms=151221485) at xfns.c:3016
609 #9 0x0815bf72 in Ffuncall (nargs=2, args=0xbf86ceec) at eval.c:2851
610
611 I installed a workaround to prevent this. The X session manager is
612 only contacted when the very first display in the Emacs session is
613 an X display. Also, x_delete_display() on this display aborts
614 session management, and XTread_socket only calls
615 x_session_check_input when it is called for the display that the
616 session was opened on. While this does not really fix the bug, it
617 makes it much less frequent, because session manager support will
618 not normally be enabled when Emacs can survive the shutdown of the
619 X server.
620
621 See if xsmfns.c should be updated.
622
623 ** Hunt down display-related functions in frame.el and extend them all
624 to accept display ids.
625
626 ** rif->flush_display_optional (NULL) calls should be replaced by a
627 new global function.
628
629 ** The set-locale-environment hack (adding the DISPLAY option) should
630 be replaced with a clean design.
631
632 ** standard-display-table should be display-local.
633 standard-display-european should be display-local.
634
635 ** With iswitchb-default-method set to 'always-frame, only frames on
636 the current display should be considered. This might involve
637 extending `get-buffer-window'.
638
639 ** Have a look at Vlocale_coding_system. Seems like it would be a
640 tedious job to localize it, although most references use it for
641 interfacing with libc and are therefore OK with the global
642 definition.
643
644 Exceptions found so far: x-select-text and
645 x-cut-buffer-or-selection-value.
646
647 ** Have a look at fatal_error_hook.
648
649 ** Have a look at set_frame_matrix_frame.
650
651 ** Check if we got term-setup-hook right.
652
653 ** I think tip_frame should be display-local.
654
655 ** Check display reference count handling in x_create_tip_frame.
656
657 ** make-frame does not correctly handle extra parameters in its
658 argument:
659
660 (frame-parameter (make-frame (list (cons 'foobar 42))) 'foobar)
661 => nil
662
663 (This is likely an error in the CVS trunk.)
664
665 ** Dan Nicolaescu suggests that -nw should be added as an alias for -t
666 in emacsclient. Good idea. (Alas, implementing this is not
667 trivial, getopt_long does not seem to support two-letter ``short''
668 options. Patches are welcome.)
669
670 ** Mark Plaksin suggests that emacsclient should accept the same
671 X-related command-line arguments as Emacs. Most of the X-related
672 argument-handling is done in Lisp, so this should be quite easy to
673 implement.
674
675 ** Gergely Nagy suggests that C-x # should only kill the current
676 frame, not any other emacsclient frame that may have the same file
677 opened for editing. I think I agree with him.
678
679 ** Very strange bug: visible-bell does not work on secondary
680 terminals in xterm and konsole. The screen does flicker a bit,
681 but it's so quick it isn't noticable.
682
683 (Update: This is probably some problem with padding or whatnot on
684 the secondary terminals.)
685
686 ** Move baud_rate to struct display.
687
688 ** Implement support for starting an interactive Emacs session without
689 an initial frame. (The user would connect to it and open frames
690 later, with emacsclient.)
691
692 ** Fix Mac support (I can't do this entirely myself). Note that the
693 current state of Mac-specific source files in the multi-tty tree
694 are not useful; before starting work on Mac support, revert to
695 pristine, pre-multi-tty versions.
696
697 ** Fix W32 support (I can't do this entirely myself). Note that the
698 current state of W32-specific source files in the multi-tty tree
699 are not useful; before starting work on W32 support, revert to
700 pristine, pre-multi-tty versions.
701
702 ** Fix DOS support (I can't do this entirely myself). Note that the
703 current state of DOS-specific source files in the multi-tty tree
704 are not useful; before starting work on DOS support, revert to
705 pristine, pre-multi-tty versions.
706
707 ** Do a grep on XXX and ?? for more issues.
708
709 ** flow-ctrl.el must be updated.
710
711 ** Fix stuff_char for multi-tty. Doesn't seem to be of high priority.
712
713 DIARY OF CHANGES
714 ----------------
715
716 (ex-TODO items with explanations.)
717
718 -- Introduce a new struct for terminal devices.
719
720 (Done, see struct tty_output. The list of members is not yet
721 complete.)
722
723 -- Change the bootstrap procedure to initialize tty_list.
724
725 (Done, but needs review.)
726
727 -- Change make-terminal-frame to support specifying another tty.
728
729 (Done, new frame parameters: `tty' and `tty-type'.)
730
731 -- Implement support for reading from multiple terminals.
732
733 (Done, read_avail_input tries to read from each terminal, until one
734 succeeds. MULTI_KBOARD is not used. Secondary terminals don't send
735 SIGIO!)
736
737 (Update: They do, now.)
738
739 (Update2: After enabling X, they don't.)
740
741 -- other-frame should cycle through the frames on the `current'
742 terminal only.
743
744 (Done, by trivially modifiying next_frame and prev_frame.)
745
746 -- Support different terminal sizes.
747
748 (Done, no problem.)
749
750 -- Make sure terminal resizes are handled gracefully. (Could be
751 problematic.)
752
753 (Done. We don't get automatic SIGWINCH for additional ttys,
754 though.)
755
756 -- Extend emacsclient to automatically open a new tty when it connects
757 to Emacs.
758
759 (Done. It's an ugly hack, needs more work.)
760
761 -- Redisplay must refresh the topmost frame on *all* terminals, not
762 just the initial terminal.
763
764 (Done, but introduced an ugly redisplay problems. Ugh.)
765
766 -- Fix redisplay problems.
767
768 (Done; it turned out that the entire Wcm structure must be moved
769 inside tty_output. Why didn't I catch this earlier?)
770
771 -- Provide a way for emacsclient to tell Emacs that the tty has been
772 resized.
773
774 (Done, simply forward the SIGWINCH signal.)
775
776 -- Each keypress should automatically select the frame corresponding
777 to the terminal that it was coming from. This means that Emacs
778 must know from which terminal the last keyboard event came from.
779
780 (Done, it was quite simple, the input event system already
781 supported multiple frames.)
782
783 -- Fix SIGIO issue with secondary terminals.
784
785 (Done, emacsclient signals Emacs after writing to the proxy pseudo
786 terminal. Note that this means that multi-tty does not work with
787 raw ttys!)
788
789 (Update: This is bullshit. There is a read_input_waiting function,
790 extend that somehow.)
791
792 (Update of update: The first update was not right either, extending
793 read_input_waiting was not necessary. Secondary ttys do seem to
794 send signals on input.)
795
796 (Update^3: Not any more.)
797
798 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
799 parameters from the currently selected terminal before the global
800 default.
801
802 (Done.)
803
804 -- Put all cached terminal escape sequences into struct tty_output.
805 Currently, they are still stored in global variables, so we don't
806 really support multiple terminal types.
807
808 (Done. It was not fun.)
809
810 -- Implement sane error handling after initialization. (Currently
811 emacs exits if you specify a bad terminal type.) The helpful error
812 messages must still be provided when Emacs starts.
813
814 (Done.)
815
816 -- Implement terminal deletion, i.e., deleting local frames, closing
817 the tty device and restoring its previous state without exiting
818 Emacs.
819
820 (Done, but at the moment only called when an error happens during
821 initialization. There is a memory corruption error around this
822 somewhere.) (Update: now it is fully enabled.)
823
824 -- Implement automatic deletion of terminals when the last frame on
825 that terminal is closed.
826
827 (Done.)
828
829 -- Restore tty screen after closing the terminal.
830
831 (Done, we do the same as Emacs 21.2 for all terminals.)
832
833 -- 'TERM=dumb src/emacs' does not restore the terminal state.
834
835 (Done.)
836
837 -- C-g should work on secondary terminals.
838
839 (Done, but the binding is not configurable.)
840
841 -- Deal with SIGHUP in Emacs and in emacsclient. (After this, the
842 server-frames may be removed from server.el.)
843
844 (Done, nothing to do. It seems that Emacs does not receive SIGHUP
845 from secondary ttys, which is actually a good thing.) (Update: I
846 think it would be a bad idea to remove server-frames.)
847
848 -- Change emacsclient/server.el to support the -t argument better,
849 i.e. automatically close the socket when the frame is closed.
850
851 (Seems to be working OK.)
852
853 -- Fix mysterious memory corruption error with tty deletion. To
854 trigger it, try the following shell command:
855
856 while true; do TERM=no-such-terminal-definition emacsclient -h; done
857
858 Emacs usually dumps core after a few dozen iterations. (The bug
859 seems to be related to the xfreeing or bzeroing of
860 tty_output.Wcm. Maybe there are outside references to struct Wcm?
861 Why were these vars collected into a struct before multi-tty
862 support?)
863
864 (Done. Whew. It turned out that the problem had nothing to do
865 with hypothetical external references to Wcm, or any other
866 tty_output component; it was simply that delete_tty closed the
867 filehandles of secondary ttys twice, resulting in fclose doubly
868 freeing memory. Utterly trivial matter. I love the C's memory
869 management, it puts hair on your chest.)
870
871 -- Support raw secondary terminals. (Note that SIGIO works only on
872 the controlling terminal.) Hint: extend read_input_waiting for
873 multiple ttys and hopefully this will be fixed.
874
875 (Done, it seems to have been working already for some time. It
876 seems F_SETOWN does work, after all. Not sure what made it fail
877 earlier, but it seems to be fixed (there were several changes
878 around request_sigio, maybe one of them did it).
879 read_input_waiting is only used in sys_select, don't change
880 it.) (Update: After adding X support, it's broken again.)
881 (Update^2: No it isn't.) :-)
882
883 -- Find out why does Emacs abort when it wants to close its
884 controlling tty. Hint: chan_process[] array. Hey, maybe
885 noninterrupt-IO would work, too? Update: no, there is no process
886 for stdin/out.
887
888 (Done. Added add/delete_keyboard_wait_descriptor to
889 term_init/delete_tty. The hint was right, in a way.)
890
891 -- Issue with SIGIO: it needs to be disabled during redisplay. See if
892 fcntl kernel behaviour could be emulated by emacsclient.
893
894 (Done. Simply disabled the SIGIO emulation hack in emacsclient.)
895 (Update: it was added back.) (Update^2: and removed again.)
896
897 -- server.el: There are issues with saving files in buffers of closed
898 clients. Try editing a file with emacsclient -f, and (without
899 saving it) do a delete-frame. The frame is closed without
900 question, and a surprising confirmation prompt appears in another
901 frame.
902
903 (Done. delete-frame now asks for confirmation if it still has
904 pending buffers, and modified buffers don't seem to be deleted.)
905
906 -- emacsclient.el, server.el: Handle eval or file open errors when
907 doing -t.
908
909 (Done.)
910
911 -- Make parts of struct tty_output accessible from Lisp. The device
912 name and the type is sufficient.
913
914 (Done, see frame-tty-name and frame-tty-type.)
915
916 -- Export delete_tty to the Lisp environment, for emacsclient.
917
918 (Done, see delete-tty.)
919
920 -- Get rid of the accessor macros in termchar.h, or define macros for
921 all members.
922
923 (Done.)
924
925 -- Move device-specific parameters (like costs) commonly used by
926 device backends to a common, device-dependent structure.
927
928 (Done. See struct display_method in termhooks.h.)
929
930 -- Fix X support.
931
932 (Done. Well, it seems to be working.)
933
934 -- Allow simultaneous X and tty frames. (Handling input could be
935 tricky. Or maybe not.)
936
937 (Done. Allowed, that is. It is currently extremely unstable, to
938 the point of being unusable. The rif variable causes constant
939 core dumps. Handling input is indeed tricky.)
940
941 -- Rewrite multi-tty input in terms of MULTI_KBOARD.
942
943 (Done. In fact, there was no need to rewrite anything, I just
944 added a kboard member to tty_display_info, and initialized the
945 frame's kboard from there.)
946
947 -- Fix rif issue with X-tty combo sessions. IMHO the best thing to do
948 is to get rid of that global variable (and use the value value in
949 display_method, which is guaranteed to be correct).
950
951 (Done, did exactly that. Core dumps during combo sessions became
952 much rarer. In fact, I have not yet met a single one.)
953
954 -- Add multi-tty support to talk.el.
955
956 (Done.)
957
958 -- Clean up the source of emacsclient. It is a mess.
959
960 (Done, eliminated stupid proxy-pty kludge.)
961
962 -- Fix faces on tty frames during X-tty combo sessions. There is an
963 init_frame_faces call in init_sys_modes, see if there is a problem
964 with it.
965
966 (Done, there was a stupid mistake in
967 Ftty_supports_face_attributes_p. Colors are broken, though.)
968
969 -- C-x 5 2, C-x 5 o, C-x 5 0 on an emacsclient frame unexpectedly
970 exits emacsclient. This is a result of trying to be clever with
971 delete-frame-functions.
972
973 (Fixed, added delete-tty-after-functions, and changed server.el to
974 use it.)
975
976 -- Something with (maybe) multi-keyboard support broke function keys
977 and arrows on ttys during X+tty combo sessions. Debug this.
978
979 (I can't reproduce it, maybe the terminal type was wrong.)
980
981 -- Fix input from raw ttys (again).
982
983 (Now it seems to work all right.)
984
985 -- During an X-tty combo session, a (message "Hello") from a tty frame
986 goes to the X frame. Fix this.
987
988 (Done. There was a safeguard against writing to the initial
989 terminal frame during bootstrap which prevented echo_area_display
990 from working correctly on a tty frame during a combo session.)
991
992 -- If there are no frames on its controlling terminal, Emacs should
993 exit if the user presses C-c there.
994
995 (Done, as far as possible. See the SIGTERM comment in
996 interrupt_signal on why this seems to be impossible to solve this
997 in general.)
998
999 -- During an X session, Emacs seems to read from stdin. Also, Emacs
1000 fails to start without a controlling tty.
1001
1002 (Fixed by replacing the troublesome termcap display with a dummy
1003 bootstrap display during bootstrap.
1004
1005 -- Do tty output through struct display, like graphical display
1006 backends.
1007
1008 (Done.)
1009
1010 -- Define an output_initial value for output_method for the initial
1011 frame that is dumped with Emacs. Checking for this frame (e.g. in
1012 cmd_error_internal) is ugly.
1013
1014 (Done, breaking interactive temacs.)
1015
1016 -- The command `emacsclient -t -e '(delete-frame)'' fails to exit.
1017
1018 (Fixed.)
1019
1020 -- frame-creation-function should always create a frame that is on the
1021 same display as the selected frame. Maybe frame-creation-function
1022 should simply be removed and make-frame changed to do the right
1023 thing.
1024
1025 (Done, with a nice hack. frame-creation-function is now frame-local.)
1026
1027 -- Fix C-g on raw ttys.
1028
1029 (Done. I disabled the interrupt/quit keys on all secondary
1030 terminals, so Emacs sees C-g as normal input. This looks like an
1031 overkill, because emacsclient has extra code to pass SIGINT to
1032 Emacs, so C-g should remain the interrupt/quit key on emacsclient
1033 frames. See the next entry why implementing this distinction would
1034 be a bad idea.)
1035
1036 -- Make sure C-g goes to the right frame with ttys. This is hard, as
1037 SIGINT doesn't have a tty parameter. :-(
1038
1039 (Done, the previous change fixes this as a pleasant side effect.)
1040
1041 -- I have seen a case when Emacs with multiple ttys fell into a loop
1042 eating 100% of CPU time. Strace showed this loop:
1043
1044 getpid() = 30284
1045 kill(30284, SIGIO) = 0
1046 --- SIGIO (I/O possible) @ 0 (0) ---
1047 ioctl(6, FIONREAD, [0]) = -1 EIO (Input/output error)
1048 ioctl(5, FIONREAD, [0]) = -1 EIO (Input/output error)
1049 ioctl(0, FIONREAD, [0]) = 0
1050 sigreturn() = ? (mask now [])
1051 gettimeofday({1072842297, 747760}, NULL) = 0
1052 gettimeofday({1072842297, 747806}, NULL) = 0
1053 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
1054 select(9, [0 3 5 6], NULL, NULL, {0, 0}) = 2 (in [5 6], left {0, 0})
1055 gettimeofday({1072842297, 748245}, NULL) = 0
1056
1057 I have seen something similar with a single X frame, but have not
1058 been able to reproduce it for debugging.
1059
1060 Update: This may have been caused by checking for nread != 0
1061 instead of nread > 0 after calling read_socket_hook in
1062 read_avail_input.
1063
1064 (Fixed. This was caused by unconditionally including stdin in
1065 input_wait_mask in init_process. The select call in
1066 wait_reading_process_input always returned immediately, indicating
1067 that there is pending input from stdin, which nobody read.
1068
1069 Note that the above strace output seems to be an unrelated but
1070 similar bug. I think that is now fixed.)
1071
1072 -- Exiting Emacs while there are emacsclient frames doesn't restore the
1073 ttys to their default states.
1074
1075 (This seems to be fixed by some previous change.)
1076
1077 -- Allow opening an X session after -nw.
1078
1079 (Done.)
1080
1081 -- Fix color handling during tty+X combo sessions. (It seems that tty
1082 sessions automatically convert the face colors to terminal colors
1083 when the face is loaded. This conversion must happen instead on
1084 the fly in write_glyphs, which might be problematic, as color
1085 approximation is currently done in lisp (term/tty-colors.el).)
1086 (Update: hm, colors seem to work fine if I start emacs with -nw and
1087 then create an X frame. Maybe it's just a small buglet somewhere.)
1088
1089 (Seems to be fixed. The problem was in startup.el, it did not
1090 initialize tty colors when the initial window system was
1091 graphical.)
1092
1093 -- emacs -nw --eval '(y-or-n-p "Foobar")' segfaults. (Reported by
1094 Romain Francoise)
1095
1096 (Fixed, there was a keyboard initialization problem.)
1097
1098 -- Fix interactive use of temacs. There are face-related SEGVs, most
1099 likely because of changes in realize_default_face, realize_face.
1100
1101 (Fixed.)
1102
1103 -- Don't exit Emacs when the last X connection fails during a
1104 multi-display session.
1105
1106 (Fixed.)
1107
1108 -- Dan Nicolaescu noticed that starting emacsclient on the same
1109 terminal device that is the controlling tty of the Emacs process
1110 gives unexpected results.
1111
1112 (Fixed.)
1113
1114 -- Istvan Marko reported that Emacs hang on ttys if it was started
1115 from a shell script.
1116
1117 (Fixed. There was a bug in the multi-tty version of
1118 narrow_foreground_group. tcsetpgrp blocks if it is called from a
1119 process that is not in the same process group as the tty.)
1120
1121 -- emacsclient -t from an Emacs term buffer does not work, complains
1122 about face problems. This can even lock up Emacs (if the recursive
1123 frame sets single_kboard). Update: the face problems are caused by
1124 bugs in term.el, not in multi-tty. The lockup is caused by
1125 single_kboard mode, and is not easily resolvable. The best thing to
1126 do is to simply refuse to create a tty frame of type `eterm'.
1127
1128 (Fixed, changed emacsclient to check for TERM=eterm. The face
1129 complaints seem to be caused by bugs in term.el; they are not
1130 related to multi-tty.)
1131
1132 -- Find out the best way to support suspending Emacs with multiple
1133 ttys. My guess: disable it on the controlling tty, but from other
1134 ttys pass it on to emacsclient somehow. (It is (I hope) trivial to
1135 extend emacsclient to handle suspend/resume. A `kill -STOP' almost
1136 works right now.)
1137
1138 (Done. I needed to play with signal handling and the server
1139 protocol a bit to make emacsclient behave as a normal UNIX program
1140 wrt foreground/background process groups.)
1141
1142 -- There is a flicker during the startup of `emacs -nw'; it's as if
1143 the terminal is initialized, reset and then initialialized again.
1144 Debug this. (Hint: narrow_foreground_group is called twice during
1145 startup.)
1146
1147 (This is gone.)
1148
1149 -- Robert Chassell has found serious copy-paste bugs with the
1150 multi-tty branch. There seem to be redisplay bugs while copying
1151 from X to a terminal frame. Copying accented characters do not
1152 work for me.
1153
1154 (Patch-124 should fix this, by changing the interprogram-*-function
1155 variables to be frame-local, as suggested by Mark Plaksin
1156 (thanks!). I think that the redisplay bugs are in fact not bugs,
1157 but delays caused by single_kboard --> perhaps MULTI_KBOARD should
1158 be removed.)
1159
1160 -- frame-creation-function was removed, which might be a bad idea.
1161 Think up a compatible solution.
1162
1163 (It was an internal interface that may be changed when necessary.)
1164
1165 -- Change Lisp code not to (getenv "TERM"); use the `tty-type' frame
1166 parameter or the frame-tty-type function instead. (M-x tags-search
1167 "TERM" helps with this.) Update: Actually, all getenv invocations
1168 should be checked for multi-tty compatibility, and an interface
1169 must be implemented to get the remote client's environment.
1170
1171 (Done. Only getenv calls in lisp/term/*.el were changed; other
1172 calls should be mostly left as they are.)
1173
1174 -- Add an elaborate mechanism for display-local variables. (There are
1175 already a few of these; search for `terminal-local' in the Elisp
1176 manual.)
1177
1178 (Not needed. Display-local variables could be emulated by
1179 frame-local variables.)
1180
1181 -- Emacs assumes that all terminal frames have the same locale
1182 settings as Emacs itself. This may lead to bogus results in a
1183 multi-locale setup. (E.g., while logging in from a remote client
1184 with a different locale.)
1185 (Update after new bugreport by Friedrich Delgado Friedrichs:
1186 (at least) the structs terminal_coding and keyboard_coding in
1187 coding.c must be moved to struct display, and the Lisp interface
1188 [set-]keyboard-coding-system must be adapted for the change.)
1189
1190 (Fixed. Emacs now uses the locale settings as seen by the
1191 emacsclient process for server tty frames.)
1192 (Update: Not really; Vlocale_coding_system is still global.)
1193
1194 -- Make `struct display' accessible to Lisp programs. Accessor functions:
1195
1196 (displayp OBJECT): Returns t if OBJECT is a display.
1197 => Implemented as display-live-p.
1198
1199 (display-list): Returns list of currently active displays.
1200 => Implemented.
1201
1202 (selected-display): Returns the display object of the selected frame.
1203 => Not strictly necessary, but implemented anyway.
1204
1205 (frame-display FRAME): Returns the display object of FRAME.
1206 => Implemented.
1207
1208 (display-frames DISPLAY): Returns a list of frames on DISPLAY.
1209 => Already implemented, see frames-on-display-list.
1210
1211 (display-type DISPLAY): Returns the type of DISPLAY, as a
1212 symbol. (See `framep'.)
1213 => Implemented as display-live-p.
1214
1215 (display-device DISPLAY): Returns the name of the device that
1216 DISPLAY uses, as a string. (E.g: "/dev/pts/16", or
1217 ":0.0")
1218 => Implemented as display-name.
1219
1220 etc.
1221
1222 See next issue why this is necessary.
1223
1224 (Update: The consensus on emacs-devel seems to be to do this via
1225 integer identifiers. That's fine by me.)
1226
1227 (Done.)
1228
1229 -- The following needs to be supported:
1230
1231 $ emacsclient -t
1232 C-z
1233 $ emacsclient -t
1234 (This fails now.)
1235
1236 The cleanest way to solve this is to allow multiple displays on the
1237 same terminal device; each new emacsclient process should create
1238 its own display. As displays are currently identified by their
1239 device names, this is not possible until struct display becomes
1240 accessible as a Lisp-level object.
1241
1242 (Done.)
1243
1244 -- Miles Bader suggests that C-x C-c on an emacsclient frame should
1245 only close the frame, not exit the entire Emacs session. Update:
1246 see above for a function that does this. Maybe this should be the
1247 new default?
1248
1249 (Done. This is the new default. No complaints so far.)
1250
1251 -- Clean up the frame-local variable system. I think it's ugly and
1252 error-prone. But maybe I just haven't yet fully understood it.
1253
1254 (Nothing to do. It doesn't seem ugly any more. It's rather clever.)
1255
1256 -- Support multiple character locales. A version of
1257 `set-locale-environment' needs to be written for setting up
1258 display-local settings on ttys. I think calling
1259 set-display-table-and-terminal-coding-system and
1260 set-keyboard-coding-system would be enough. The language
1261 environment itself should remain a global setting.
1262
1263 (Done, by an ugly hack.)
1264
1265 -- The terminal customization files in term/*.el tend to change global
1266 parameters, which may confuse Emacs with multiple displays. Change
1267 them to tweak only frame-local settings, if possible. (They tend
1268 to call define-key to set function key sequences a lot.)
1269
1270 (Done, by making `function-key-map' terminal-local (i.e., part of
1271 struct kboard). This has probably covered all the remaining problems.)
1272
1273 -- Make `function-key-map' and `key-translation-map' terminal-local.
1274
1275 (Done.)
1276
1277 -- Implement `terminal-local-value' and `set-terminal-local-value' to
1278 allow deterministic access to terminal local bindings. The
1279 encode-kb package can not set up `key-translation-map' without
1280 these. The terminal-local bindings seem to be independent of what
1281 frame is selected.
1282
1283 (Done.)
1284
1285 -- xt-mouse.el needs to be adapted for multi-tty. It currently
1286 signals an error on kill-emacs under X, which prevents the user
1287 from exiting Emacs. (Reported by Mnemonikk on freenode.)
1288
1289 (Done, I hope.)
1290
1291
1292 -- Having {reset,init}_all_sys_modes in set-input-mode breaks arrow
1293 keys on non-selected terminals under screen, and sometimes on other
1294 terminal types as well. The other function keys continue to work
1295 fine. Sometimes faces on these screens become garbled.
1296
1297 This only seems to affect displays that are of the same terminfo
1298 type as the selected one. Interestingly, in screen Emacs normally
1299 reports the up arrow key as `M-o A', but after the above SNAFU, it
1300 complains about `M-[ a'. UNIX ttys are a complete mystery to me,
1301 but it seems the reset-reinitialize cycle somehow leaves the
1302 non-selected terminals in a different state than usual. I have no
1303 idea how this could happen.
1304
1305 Currently set-input-mode resets only the currently selected
1306 terminal, which seems to somehow work around the problem.
1307
1308 Update:
1309
1310 Dan Nicolaescu <dann@ics.uci.edu> writes:
1311 > Some terminals have 2 modes for cursor keys: Application Mode where
1312 > the cursor keys transmit the codes defined in the terminfo entry, and
1313 > Cursor mode. Applications have to send the smkx and rmkx terminfo
1314 > strings to switch between the 2 modes. So Emacs (and emacsclient) have
1315 > to send smkx when initializing and rmkx when quitting (or on
1316 > suspend).
1317
1318 (I think patch-370 fixed this.)
1319
1320 -- This long-standing bug (first reported by Han Boetes) seems to come
1321 and go all the time. It is time to track it down and fix it.
1322
1323 emacs
1324 M-x server-start
1325
1326 # From another xterm:
1327 emacsclient -e '(y-or-n-p "Do you want me to crash? ")'
1328 # Notice how the answer ends up in the *scratch* buffer
1329 M-x garbage-collect
1330 SIGSEGV
1331
1332 (Fixed in patch-414 after detailed analysis by Kalle Olavi Niemitalo.)
1333
1334 -- normal-erase-is-backspace-mode in simple.el needs to be updated for
1335 multi-tty (rep. by Dan Waber). (The Delete key is broken on X
1336 because of this.)
1337
1338 (Fixed in patch-427.)
1339
1340 -- I think keyboard-translate-table should be made terminal-local.
1341
1342 (Done in patch-431.)
1343
1344 -- The semantics of terminal-local variables are confusing; it is not
1345 clear what binding is in effect in any given time. See if
1346 current_kboard (or at least the terminal-local bindings exported to
1347 Lisp) might be changed to be tied to the selected frame instead.
1348 Currently, `function-key-map' and `key-translation-map' may be
1349 accessed reliably only using the hackish
1350 `(set-)terminal-local-value' functions.
1351
1352 Perhaps there should be a difference between `last-command' &co.
1353 and these more conventional configuration variables.
1354 (E.g. `symbol-value' would use current_kboard to access
1355 `last-command', but SELECTED_FRAME()->display->kboard to get the
1356 value of `function-key-map'.
1357
1358 (Fixed in patch-434.)
1359
1360 -- If the first key pressed on a new tty terminal is a function key,
1361 it is not recognized correctly. May be related to the bug below.
1362
1363 (Seems to have been fixed as a side effect of patch-434. "The bug
1364 below" was the set-input-mode madness.)
1365
1366 (Update: this bug was fixed for good in patch-449. It was tracked
1367 down to a bug in `read_key_sequence': it failed to reinitialize its
1368 local function-key-map/key-translation-map references when it
1369 switched keyboards. I don't understand why did this bug only
1370 appear on brand new frames, though!)
1371
1372 -- Disable connecting to a new X display when we use the GTK toolkit.
1373
1374 (Disabled in patch-450.)
1375
1376 -- Implement automatic forwarding of client environment variables to
1377 forked processes, as discussed on the multi-tty list. Terminal
1378 parameters are now accessible in C code, so the biggest obstacle is
1379 gone. The `getenv_internal' and `child_setup' functions in
1380 callproc.c must be changed to support the following variable:
1381
1382 terminal-local-environment-variables is a variable defined in ...
1383
1384 Enable or disable terminal-local environment variables.
1385
1386 If set to t, `getenv', `setenv' and subprocess creation
1387 functions use the environment variables of the emacsclient
1388 process that created the selected frame, ignoring
1389 `process-environment'.
1390
1391 If set to nil, Emacs uses `process-environment' and ignores
1392 the client environment.
1393
1394 Otherwise, `terminal-local-environment-variables' should be a
1395 list of variable names (represented by Lisp strings) to look
1396 up in the client environment. The rest will come from
1397 `process-environment'.
1398
1399 (Implemented in patch-461; `terminal-getenv', `terminal-setenv' and
1400 `with-terminal-environment' are now replaced by extensions to
1401 `getenv' and `setenv', and the new `local-environment-variables'
1402 facility. Yay!)
1403
1404 (Updated in patch-465 to fix the semantics of let-binding
1405 `process-environment'. `process-environment' was changed to
1406 override all local/global environment variables, and a new variable
1407 `global-environment' was introduced to have `process-environment's
1408 old meaning.)
1409
1410 (Updated in patch-466 to fix the case when two emacsclient sessions
1411 share the same terminal, but have different environment. The local
1412 environment lists are now stored as frame parameters, so the
1413 C-level terminal parameters are not strictly necessary any more.)
1414
1415 -- `Fdelete_frame' is called from various critical places where it is
1416 not acceptable for the frame deletion to fail, e.g. from
1417 x_connection_closed after an X error. `Fdelete_frame' now protects
1418 against `delete-frame-functions' throwing an error and preventing a
1419 frame delete. (patch-475)
1420
1421 -- Fix set-input-mode for multi-tty. It's a truly horrible interface;
1422 what if we'd blow it up into several separate functions (with a
1423 compatibility definition)?
1424
1425 (Done. See `set-input-interrupt-mode', `set-output-flow-control',
1426 `set-input-meta-mode' and `set-quit-char'.) (patch-457)
1427
1428 -- Let-binding `overriding-terminal-local-map' on a brand new frame
1429 does not seem to work correctly. (See `fancy-splash-screens'.)
1430 The keymap seems to be set up right, but events go to another
1431 terminal. Or is it `unread-command-events' that gets Emacs
1432 confused? Investigate.
1433
1434 (Emacs was confused because a process filter entered
1435 `recursive-edit' while Emacs was reading input. I added support
1436 for this in the input system.) (patch-489)
1437
1438 -- I smell something funny around pop_kboard's "deleted kboard" case.
1439 Determine what are the circumstances of this case, and fix any
1440 bug that comes to light.
1441
1442 (It happens simply because single_kboard's terminal is sometimes
1443 deleted while executing a command on it, for example the one that
1444 kills the terminal. There was no bug here, but I rewrote the whole
1445 single_kboard mess anyway.) (patch-489)
1446
1447 -- Understand Emacs's low-level input system (it's black magic) :-)
1448 What exactly does interrupt_input do? I tried to disable it for
1449 raw secondary tty support, but it does not seem to do anything
1450 useful. (Update: Look again. X unconditionally enables this, maybe
1451 that's why raw terminal support is broken again. I really do need
1452 to understand input.)
1453 (Update: I am starting to understand the read_key_sequence->read-char
1454 ->kbd_buffer_get_event->read_avail_input->read_socket_hook path. Yay!)
1455
1456 (Update: OK, it all seems so easy now (NOT). Input could be done
1457 synchronously (with wait_reading_process_input), or asynchronously
1458 by SIGIO or polling (SIGALRM). C-g either sets the Vquit_flag,
1459 signals a 'quit condition (when immediate_quit), or throws to
1460 `getcjmp' when Emacs was waiting for input when the C-g event
1461 arrived.)
1462
1463 -- Replace wrong_kboard_jmpbuf with a special return value of
1464 read_char. It is absurd that we use setjmp/longjmp just to return
1465 to the immediate caller.
1466
1467 (Done in patch-500.)
1468
1469 ;;; arch-tag: 8da1619e-2e79-41a8-9ac9-a0485daad17d
1470