aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
* If a client is killed while suspended with ^Z so has gone through thenicm2014-07-13
| | | | | MSG_EXITED dance, don't try to resume it since a) it's pointless and b) the tty structures have been cleaned up and tmux will crash.
* Track whether a process is a zombie or not yet fully built via flagsguenther2014-07-04
| | | | | | | | | | | PS_{ZOMBIE,EMBRYO} on the process instead of peeking into the process's thread data. This eliminates the need for the thread-level SDEAD state. Change kvm_getprocs() (both the sysctl() and kvm backends) to report the "most active" scheduler state for the process's threads. tweaks kettenis@ feedback and ok matthew@
* Handle escaped back slashes and missing new lines at end of line by usingtobias2014-06-25
| | | | | | fparseln instead of reimplementing it on our own. with input by and ok nicm@
* Allow keys and send-keys to invisible panes, from saggy-kun at users dotnicm2014-06-23
| | | | sf dot net.
* Count mouse clicks correctly, from Balazs Kezes.nicm2014-06-23
|
* Comment style nits.nicm2014-06-20
|
* Make -S- and -E- mean the start and end to capture-pane to avoid havingnicm2014-06-20
| | | | to faff around with huge numbers to get everything.
* Copy newline when at EOL in vi(1) mode, from Balazs Kezes.nicm2014-06-19
|
* Reset the buttons when the wheel is used, from Balazs Kezes.nicm2014-06-19
|
* Some terminals send spurious releases for mouse wheel in SGR mouse mode,nicm2014-06-19
| | | | | | this causes confusion when tmux uses SGR outside but the application inside tmux is using conventional xterm mouse reporting. So suppress obviously bad input. From Timothy Allen, SF bug 128.
* Properly track switching G0 and G1 modes between US-ASCII and VT100 linenicm2014-06-06
| | | | drawing rather than just treating them as SO and SI.
* Reset properly when c0-change-trigger is increased from zero so panesnicm2014-06-05
| | | | don't get stuck.
* Handle the top bit of xterm(1)-style modifier keys, based on a diff fromnicm2014-05-27
| | | | Balazs Kezes.
* Add some formats for pane bounds.nicm2014-05-27
|
* Change key-table to mode-table to allow for some future work. From Keithnicm2014-05-14
| | | | Amling.
* Simplify copy lines, from Keith Amling.nicm2014-05-14
|
* Now that cmdlists are reference counted, there is no need for two-stepnicm2014-05-14
| | | | deletion via the dead_key_bindings tree. From Keith Amling.
* Don't allow multiple buffers with the same name, from Thomas Adam.nicm2014-05-13
|
* If multiple arguments are given to new-session, new-window,nicm2014-05-13
| | | | | | split-window, respawn-window or respawn-pane, pass them directly to execvp() to help avoid quoting problems. One argument still goes to "sh -c" like before. Requested by many over the years. Patch from J Raynor.
* Add a copy mode key binding to copy to a named buffer. From J Raynor.nicm2014-05-13
|
* Add support for named buffers. If you don't name a buffer, things worknicm2014-05-13
| | | | | | | | much as before - buffers are automatically named "buffer0000", "buffer0001" and so on and ordered as a stack. Buffers can be named explicitly when creating ("loadb -b foo" etc) or renamed ("setb -b buffer0000 -n foo"). If buffers are named explicitly, they are not deleted when buffer-limit is reached. Diff from J Raynor.
* Just use char ** for argv like normal people, not char *const *.nicm2014-05-09
|
* Send up and down keys for mouse wheel in alternate screen mode (when itnicm2014-05-08
| | | | normally does nothing), from Marcel Partap.
* Instead of forcing mouse scroll to 1 in choose mode, scale it downnicm2014-05-08
| | | | | instead. Means modifier keys still increase the line count, just not as much. Based on a diff from Marcel Partap.
* Plug a memory leak, from J Raynor.nicm2014-05-08
|
* Both the two previous ways of navigating panes by direction havenicm2014-05-08
| | | | | | | | | | | | | | | | irritating flaws: a) The old way of always using the top or left if the choice is ambiguous is annoying when the layout is unbalanced. b) The new way of remembering the last used pane is annoying if the layout is balanced and the leftmost is obvious to the user (because clearly if we go right from the top-left in a tiled set of four we want to end up in top-right, even if we were last using the bottom-right). So instead, use a combination of both: if there is only one possible pane alongside the current pane, move to it, otherwise choose the most recently used of the choice.
* Handle colour 8 properly in the 256 colour palette, from Timothy Allen.nicm2014-05-08
|
* fcntl.h is still needed here.nicm2014-04-29
|
* specifying ECHOCTL once is enoughjsg2014-04-25
| | | | ok nicm@
* There is no longer a need for a paste_stack struct or for global_buffersnicm2014-04-24
| | | | to be global. Move to paste.c.
* Differentiate between linked and unlinked window closes and renames,nicm2014-04-23
| | | | like we already do for adds. From Andre Masella.
* #nnT went away a while ago, remove a leftover from the manpage.nicm2014-04-17
|
* Add some UTF-8 utility functions and use them to prevent the width limitnicm2014-04-17
| | | | on formats from splitting UTF-8 characters improperly.
* Some more long lines.nicm2014-04-17
|
* Only scroll by one line at a time in choose mode, lists are generallynicm2014-04-17
| | | | pretty small.
* Set PATH explicitly, either from client or sessionnicm2014-04-17
| | | | | environment. Previously it came from the session environment. From J Raynor.
* Wrap some long lines.nicm2014-04-17
|
* Don't limit the DCS buffer to 256 bytes, expand it as needed. Requestednicm2014-04-17
| | | | by Suraj Kurapati.
* Remove some unnecessary includes and fix a typo.nicm2014-04-17
|
* Correct the dance to fix the active pane in join-pane by pulling thenicm2014-04-17
| | | | | (right) code from break-pane and window_remove_pane into a helper function.
* Remove the "info" message mechanism, this was only used for about fivenicm2014-04-17
| | | | | | mostly useless and annoying messages. Change those commands to silence on success like all the others. Still accept the -q command line flag and "quiet" server option for now.
* Extend the -q flag to set-option to suppress errors about unknownnicm2014-04-17
| | | | options - this will allow options to be removed more easily.
* Do not show the -fg, -bg and -attr options. If asked for one explicitly,nicm2014-04-17
| | | | show the equivalent -style option instead.
* Remove the monitor-content option and associated bits and bobs. It'snicm2014-04-17
| | | | | never worked very well. If there is a big demand for it to return, will consider better ways to do it.
* Memory leak in error path and unnecessary assignment, from clang.nicm2014-04-16
|
* Remove a leftover prototype and fix some spacing.nicm2014-04-16
|
* Remove the choose-list command to prepare for some later choose-* work.nicm2014-04-16
|
* Because we pass the file descriptor from client to server, tmux can'tnicm2014-04-16
| | | | usefully work if stdin is /dev/tty. Complain about it more clearly.
* Don't leak socketpair file descriptors if fork fails. Spotted by Balazsnicm2014-04-14
| | | | Kezes.
* Don't blindly increase offsets by the return value of snprintf, if therenicm2014-04-11
| | | | | wasn't enough space this will go off the end. Instead clamp to the available space. Fixes crash reported by Julien Rebetez.