Bugfix festival.

lib-src/emacsclient.c (main_argc, main_argv): New variables.
(main): Initialize them.
(fail): Use them.
(window_change, copy_from_to): Don't kill if emacs_pid is zero.
(pty_conversation): Watch the command socket, too.  Read emacs_pid
here.  Emacs and emacsclient could deadlock if Emacs tried to do a
reset_sys_modes before sending its pid.

lisp/server.el: Automatically delete the client frame when done editing.
(server-frames): New variable.
(server-process-filter, server-sentinel, server-buffer-done): Use it.
(server-process-filter): Do a redisplay before evaluating other
parameters.  (Prevents "emacsclient -h -e '(delete-frame)'" from
messing up the system.

src/dispextern.h: Update prototypes.

src/dispnew.c (window_change_signal): Do nothing if !term_initted.
(init_display): Set the frame size from the tty data after term_init.

src/emacs.c (main): Make sure things that init_sys_modes needs are 
initialized before init_display (which calls init_sys_modes now).
(sort_args): Use xfree, not free.
(shut_down_emacs) [!EMACS_HAVE_TTY_PGRP]: Use reset_all_sys_modes
instead of reset_sys_modes.

src/frame.c (make_terminal_frame): Sigh. Move terminal initialization
back to the middle of frame setup.  Handle errors by making sure that
the delete_tty() called from term_init() will see and delete this
frame.
(Fdelete_frame): Kill the frame before calling delete_tty().  Fix
condition for tty deletion.

src/keyboard.c (Fset_input_mode): Use reset_sys_modes on the current
terminal only.

src/lisp.h: Remove duplicate prototypes.

src/msdos.c (croak): use reset_all_sys_modes().

src/sysdeps.c (init_baud_rate): Added tty parameter, use it instead of CURTTY.
(child_setup_tty): Reset sigio on stdin, not CURTTY().
(reset_sigio): Added fd parameter, put explicit fcntl there.
(request_sigio, unrequest_sigio)[FASYNC]: Simply block/unblock the
SIGIO signal, don't touch the file params.  There are multiple ttys
now, and we can't disable the SIGIO from emacsclient.
(get_tty_size)[VMS]: Use tty_out instead of CURTTY().
(reset_sys_modes): Don't call cursor_to, clear_end_of_line; call
cmgoto and tty_clear_end_of_line instead.  The frame may already be
dead.  Updated reset_sigio call.

src/term.c (clear_and_of_line): Separate tty-dependent stuff to
tty_clear_end_of_line() for reset_sys_modes.
(tty_clear_end_of_line): New function.
(term_init): Added frame parameter, don't use selected_frame.
Set the frame's output_data.tty value (in case there is an error
later).  Set the frame size in Wcm, not in the frame.  Only free the
termcap buffer if there is a termcap-related error.  Call
init_sys_modes last, not first.
(deleting_tty): New variable.
(delete_tty): Use it for handling recursive calls.  Free deleted tty,
except its Wcm (there is still a dangling reference somewhere).

git-archimport-id: lorentey@elte.hu--2004/emacs--multi-tty--0--patch-19

1 files changed, 22 insertions, 13 deletions

diff --git a/README.multi-tty b/README.multi-tty
index 5abb9fafd15..ffa4b18ba81 100644
--- a/README.multi-tty
+++ b/README.multi-tty
@@ -67,11 +67,7 @@ For the NEWS file:
    specify a terminal device (`tty' parameter) and a terminal type
    (`tty-type' parameter) to `make-terminal-frame'. `tty' must be a
    terminal device created by the updated emacsclient, or there will
-   be problems with terminal input and window resizes.  (The kernel
-   notifies processes about pending input or terminal resizes only on
-   the controlling terminal, so we need emacsclient to sit on the real
-   terminal device, create SIGIO signals upon terminal input, and
-   forward SIGWINCH signals to us.)
+   be problems with terminal input and window resizes.  
 
    You can test for the presence of multiple terminal support by
    testing for the `multi-tty' feature.
@@ -160,6 +156,9 @@ DIARY OF CHANGES
    terminal.  Note that this means that multi-tty does not work with
    raw ttys!)
 
+   (Update: This is bullshit.  There is a read_input_waiting function,
+   extend that somehow.)
+
 -- Make make-terminal-frame look up the `tty' and `tty-type' frame
    parameters from the currently selected terminal before the global
    default.
@@ -191,6 +190,13 @@ DIARY OF CHANGES
 
    (Done.)
 
+-- Restore tty screen after closing the terminal.
+
+   (Done, we do the same as Emacs 21.2 for all terminals.)
+
+-- 'TERM=dumb src/emacs' does not restore the terminal state.
+   
+   (Done.)
 
 THINGS TO DO
 ------------
@@ -216,35 +222,38 @@ THINGS TO DO
 
 ** Export delete_tty to the Lisp environment, for emacsclient.
 
-** Restore tty screen after closing the terminal.
-
-** 'TERM=dumb src/emacs' does not restore the terminal state.
-
 ** C-g should work on secondary terminals.
 
 ** Make parts of struct tty_output accessible from Lisp.  The device
    name and the type is sufficient.
 
 ** Find out why does Emacs abort when it wants to close its
-   controlling tty.
+   controlling tty.  Hint: chan_process[] array.  Hey, maybe
+   noninterrupt-IO would work, too?
 
 ** Implement support for starting an interactive Emacs session without
    an initial frame.  (The user would connect to it and open frames
    later, with emacsclient.)  Not necessarily a good idea.
 
-** Support raw secondary terminals.  (This one is tricky, SIGIO works
-   only on the controlling terminal.  The emacsclient solution works
-   nicely, so this is not that important anyway.)
+** Support raw secondary terminals.  (Note that SIGIO works only on
+   the controlling terminal.) Hint: extend read_input_waiting() for
+   multiple ttys and hopefully this will be fixed.
 
 ** What does interrupt_input do?  I tried to disable it for raw
    secondary tty support, but it does not seem to do anything useful.
 
+** Issue with SIGIO: it needs to be disabled during redisplay.  See if
+   fcntl() kernel behaviour could be emulated by emacsclient.
+
 ** Move optimalization parameters (costs) from union output_data to
    a backend-neutral per-device structure.
 
 ** Find out the best way to support suspending Emacs with multiple
    ttys.
 
+** Deal with SIGHUP in Emacs and in emacsclient.  (After this, the
+   server-frames may be removed from server.el.)
+
 ** Do tty output through term_hooks, like all other display backends.
 
 ** Fix X support.