Call bufferevent_free before closing file descriptor associated with it
or bugs in $EventMechanism on $OtherOS makes libevent get it's knickers
in a twist. From Dylan Alex Simon.
Don't reset the activity timer for unattached sessions every second,
this screws up the choice of most-recently-used. Instead, break the time
update into a little function and do it when the session is attached.
Pointed out by joshe@.
this screws up the choice of most-recently-used. Instead, break the time
update into a little function and do it when the session is attached.
Pointed out by joshe@.
Store sessions in an RB tree by name rather than a list, this is tidier
and allows them to easily be shown sorted in various lists
(list-sessions/choose-sessions).
Keep a session index which is used in a couple of places internally but
make it an ever-increasing number rather than filling in gaps with new
sessions.
and allows them to easily be shown sorted in various lists
(list-sessions/choose-sessions).
Keep a session index which is used in a couple of places internally but
make it an ever-increasing number rather than filling in gaps with new
sessions.
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.
- 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.
When changing so that the client passes its stdout and stderr as well as
stdin up to the server, I forgot one essential point - the tmux server
could now be both the producer and consumer. This happens when tmux is
run inside tmux, as well as when piping tmux commands together.
So, using stdio(3) was a bad idea - if sufficient data was written, this
could block in write(2). When that happened and the server was both
producer and consumer, it deadlocks.
Change to use libevent bufferevents for the client stdin, stdout and
stderr instead. This is trivial enough for output but requires a
callback mechanism to trigger when stdin is finished.
This relies on the underlying polling mechanism for libevent to work
with whatever devices to which the user could redirect stdin, stdout or
stderr, hence the change to use poll(2) over kqueue(2) for tmux.
stdin up to the server, I forgot one essential point - the tmux server
could now be both the producer and consumer. This happens when tmux is
run inside tmux, as well as when piping tmux commands together.
So, using stdio(3) was a bad idea - if sufficient data was written, this
could block in write(2). When that happened and the server was both
producer and consumer, it deadlocks.
Change to use libevent bufferevents for the client stdin, stdout and
stderr instead. This is trivial enough for output but requires a
callback mechanism to trigger when stdin is finished.
This relies on the underlying polling mechanism for libevent to work
with whatever devices to which the user could redirect stdin, stdout or
stderr, hence the change to use poll(2) over kqueue(2) for tmux.
New option, detach-on-destroy, to set what happens to a client when the session
it is attached to is destroyed. If on (the default), it is detached; if off, it
is switched to the most recently active session.
If remain-on-exit is set, both the error callback and a SIGCHLD could
destroy the same pane (because the first one doesn't remove it from the
list of panes), causing the pane bufferevent to be freed twice. So don't
free it if the fd has already been set to -1, from Romain Francoise.
destroy the same pane (because the first one doesn't remove it from the
list of panes), causing the pane bufferevent to be freed twice. So don't
free it if the fd has already been set to -1, from Romain Francoise.
