Commit Graph

94 Commits

Author SHA1 Message Date
Nicholas Marriott
9a7cde0c9b Two new options:
- server option "exit-unattached" makes the server exit when no clients
  are attached, even if sessions are present;

- session option "destroy-unattached" destroys a session once no clients
  are attached to it.

These are useful for preventing tmux remaining in the background where
it is undesirable and when using tmux as a login shell to keep a limit
on new sessions.
2010-09-26 20:43:30 +00:00
Nicholas Marriott
66152010a7 Modify the permissions on the socket when adding or removing +x to show
attached sessions, rather than replacing them.
2010-09-26 18:51:48 +00:00
Nicholas Marriott
e3be9b1951 Do not call event_del() for signals after fork(), just use sigaction()
directly instead - calling libevent functions after fork() w/o
event_reinit() is a bad idea, even if in this case it was harmless.
2010-08-19 18:29:01 +00:00
Theo Deraadt
4274a7ec89 switch back to kqueue for now, since (a) kqueue has been fixed to deal
with strange devices and (b) since there appears to be a bull in the
poll code in libevent as well...
requested by nicm who is away
2010-08-04 19:46:13 +00:00
Nicholas Marriott
bf09b00fe9 kqueue(2) is currently broken when used with /dev/null and a few other
devices.

An upcoming fix for some problems with the client stdout/stderr handling
relies on it working, so make tmux force libevent to use poll(2) via
EVENT_NOKQUEUE, until we have fixed kqueue.
2010-07-24 19:25:31 +00:00
Nicholas Marriott
ad466a69aa Give tmux sockets (but not the containing folder) group
permissions. This allows hardlinks to the sockets to be used more
easily.
2010-06-21 00:18:57 +00:00
Nicholas Marriott
c4a2fdf15b Put this back in with the initialisation in the right order. 2010-05-04 17:28:16 +00:00
Nicholas Marriott
af5e0bd15a Revert last change, it appears to be broken somehow. 2010-05-04 08:48:06 +00:00
Nicholas Marriott
ec1d37b1b2 Make signal handler setup/teardown two common functions instead of six,
and reset SIGCHLD after fork to fix problems with some shells. From
Romain Francois.
2010-05-03 16:06:32 +00:00
Nicholas Marriott
ac9daf92d7 Merge copy mode and output mode, dropping the latter. Idea and code from
Micah Cowan.
2010-04-06 21:35:44 +00:00
Nicholas Marriott
509ce7f766 Nuke unused variable. 2010-03-22 19:11:54 +00:00
Nicholas Marriott
4651180503 Use the array.h code for the causes list. 2010-02-06 23:22:27 +00:00
Nicholas Marriott
8aba77b7be Instead of bailing out on the first configuration file error, carry on,
collecting all the errors, then start with the active window in more mode
displaying them.
2010-02-06 17:15:33 +00:00
Nicholas Marriott
65c9004550 Ignore SIGHUP as well. 2010-01-30 19:05:18 +00:00
Nicholas Marriott
15a64b805e Massive spaces->tabs and trailing whitespace cleanup, hopefully for the last
time now I've configured emacs to make them displayed in really annoying
colours...
2009-12-03 22:50:09 +00:00
Nicholas Marriott
459e9de81a Close the pane if the process died due to a signal, not just if it exited
normally.
2009-12-02 15:06:14 +00:00
Nicholas Marriott
4ca857e0e9 Remove a couple of unused arguments where possible, and add /* ARGSUSED */ to
the rest to reduce lint output.
2009-11-26 21:37:13 +00:00
Nicholas Marriott
8e47966225 Destroy panes immediately rather than checking them all every loop. 2009-11-13 17:33:07 +00:00
Nicholas Marriott
08a8ccf46b Free the pane bufferevent when the fd is closed (the signal could come before
the error callback).
2009-11-11 13:24:42 +00:00
Nicholas Marriott
2756437f4b Only need to chmod +x or -x the socket when a client is created, lost or
attached, rather than every event loop.
2009-11-11 08:00:42 +00:00
Nicholas Marriott
25c604fb1c EVLOOP_ONCE takes care of the wakeup, so no need to call event_loopexit(NULL). 2009-11-05 08:50:32 +00:00
Nicholas Marriott
38e13942ac Now all timers are events, there is no longer any need to wake up every 50 ms -
only wake up when an event happens.
2009-11-05 08:48:15 +00:00
Nicholas Marriott
946ed97273 Move status timer check into the global once-per-second timer, this could maybe
be done better but one every second is better than once every 50 ms.
2009-11-04 23:42:51 +00:00
Nicholas Marriott
b3c4956efe Don't reenlist the client imsg event every loop, instead have a small function
to it and call it after the event triggers or after a imsg is added.
2009-11-04 23:12:43 +00:00
Nicholas Marriott
fde36fccc3 Tell the client to exit on configuration file error. 2009-11-04 22:47:34 +00:00
Nicholas Marriott
a02c7e804c Convert the window pane (pty master side) fd over to use a bufferevent.
The evbuffer API is very similar to the existing tmux buffer API so this was
remarkably painless. Not many possible ways to do it, I suppose.
2009-11-04 22:43:11 +00:00
Nicholas Marriott
06ffed3216 Call event_init() before loading the config file, since potentially it could
set up events.
2009-11-04 22:40:36 +00:00
Nicholas Marriott
4d6091379b Switch jobs over to use a bufferevent. 2009-11-04 21:04:43 +00:00
Nicholas Marriott
abf3a5d50e Initial changes to move tmux to libevent.
This moves the client-side loops are pretty much fully over to event-based only
(tmux.c and client.c) but server-side (server.c and friends) treats libevent as
a sort of clever poll, waking up after every event to run various things.

Moving the server stuff over to bufferevents and timers and so on will come
later.
2009-11-04 20:50:11 +00:00
Nicholas Marriott
5289da29ba Change session and client activity and creation time members to have more
meaningful names.

Also, remove the code to try and update the session activity time for the
command client when a command message is received as is pointless because it
des not have a session.
2009-11-03 20:29:47 +00:00
Nicholas Marriott
992dd86309 Reorder slightly to tidy code. 2009-11-02 12:48:44 +00:00
Nicholas Marriott
37ffdff5ba Move the poll registration functions into the server-*.c files. 2009-10-27 13:03:33 +00:00
Theo Deraadt
ed62d1263c tabs are better; ok nicm 2009-10-26 21:42:04 +00:00
Nicholas Marriott
a8b1379ccb Clear signal flags /before/ taking action and continue afterwards to reduce
chance of dropping signals. Pointed out by deraadt@.
2009-10-26 21:38:18 +00:00
Nicholas Marriott
eddcc3dfa9 Split the server code handling clients, jobs and windows off into separate
files from server.c (merging server-msg.c into the client file) and rather than
iterating over each set after poll(), allow a callback to be specified when the
fd is added and just walk once over the returned pollfds calling each callback
where needed.

More to come, getting this in so it is tested.
2009-10-22 19:41:51 +00:00
Nicholas Marriott
9294cb099f When a session is unattached, reset its activity timer to prevent it locking
instantly when reattached.
2009-10-13 06:14:08 +00:00
Nicholas Marriott
762459954f Similarly add a tty_cursor_pane function to tidy up most of the calls. 2009-10-12 09:29:58 +00:00
Nicholas Marriott
972a6f5656 _absolute is redundant, just use tty_region. 2009-10-12 09:16:59 +00:00
Nicholas Marriott
f05b32f7ad Cleanup: use two functions for region setting, one for absolute and one inside
pane.
2009-10-12 09:09:35 +00:00
Nicholas Marriott
64b5f39656 Add a pipe-pane command to allow a pane to be piped to a shell command, for
example:

	pipe-pane 'cat >~/out'

No arguments stops outputing and closes the pipe; the -o flag toggles a pipe
and on and off (useful for key bindings).

Suggested by espie@.
2009-10-11 10:04:27 +00:00
Nicholas Marriott
cebc988dd4 Switch run-shell over to queue the command in the background like #(). 2009-10-11 08:58:05 +00:00
Nicholas Marriott
ff4b4e667a Collect status from dead jobs and don't invoke the callback until both
all input (the socket is closed) and status is available.
2009-10-11 07:20:16 +00:00
Nicholas Marriott
4bc0f6e7e9 Clean up by introducing a wrapper struct for mouse clicks rather than passing
three u_chars around.

As a side-effect this fixes incorrectly rejecting high cursor positions
(because it was comparing them as signed char), reported by Tom Doherty.
2009-10-11 07:01:10 +00:00
Nicholas Marriott
f68ade7b1d Braek some bits out of server_fill_client() that aren't really related to
polling into their own function.
2009-10-11 00:53:14 +00:00
Nicholas Marriott
095ecf2d90 Put all jobs on a global all_jobs list and use that in server.c instead of
running through all the clients.
2009-10-10 18:42:14 +00:00
Nicholas Marriott
6bca92db4d Rather than running status-left, status-right and window title #() with popen
immediately every redraw, queue them up and run them in the background,
starting each once every status-interval. The actual status line uses the
output from the last run.

This brings several advantages:

- tmux itself may be called from inside #() without causing the server to hang;
- likewise, sleep or similar doesn't cause the server to block;
- commands aren't run excessively often when redrawing;
- commands shared by status-left and status-right, or used multiple times, will
  only be run once.

run-shell and if-shell still use system()/popen() but will be changed over to
use this too later.
2009-10-10 15:03:01 +00:00
Nicholas Marriott
4658c063d5 New option, mouse-select-pane. If on, the mouse may be used to select the
current pane.

Suggested by sthen@ and also by someone else ages ago who I have forgotten.
2009-10-10 14:51:16 +00:00
Nicholas Marriott
9dd72b9583 Add "grouped sessions" which have independent name, options, current window and
so on but where the linked windows are synchronized (ie creating, killing
windows and so on are mirrored between the sessions). A grouped session may be
created by passing -t to new-session.

Had this around for a while, tested by a couple of people.
2009-10-10 10:02:48 +00:00
Nicholas Marriott
b7d031cc92 Support for individual session idle time locking. May be enabled by turning off
the lock-server option (it is on by default). When this is off, each session
locks when it has been idle for the lock-after-time setting. When on, the
entire server locks when ALL sessions have been idle for their individual
lock-after-time settings.

This replaces one global-only option (lock-after-time) with another
(lock-server), but the default behaviour is usually preferable so there don't
seem to be many alternatives.

Diff/idea largely from Thomas Adam, tweaked by me.
2009-10-10 09:46:11 +00:00
Nicholas Marriott
93b353d353 Instead of passing a struct pollfd ** around through various functions, build
them into a tree and then convert into a flat poll array before and after poll.

This adds a little code but should reduce annoying problems with ordering when
adding new things that also need to be polled.
2009-10-10 09:31:39 +00:00