Commit Graph

180 Commits

Author SHA1 Message Date
nicm
43264dfbf4 Make the mode draw function use the parent screen directly rather than
its own to avoid copying twice.
2017-11-02 22:00:42 +00:00
nicm
8d37f699ad Add a "fast" version of screen_write_copy for tree mode that doesn't do
all the checks and selection and marking stuff needed for copy mode.
2017-11-02 21:29:17 +00:00
nicm
6a292f09ba When writing batches of characters to the screen, we need to clear
padding or later UTF-8 characters could be displayed incorrectly. GitHub
issue 1090.
2017-10-05 08:12:24 +00:00
nicm
5dd5543fe4 Add -F to choose-tree, choose-client, choose-buffer to specify the
format of each line, as well as adding a couple of formats needed for
the default display.
2017-08-09 11:43:45 +00:00
nicm
7247553c77 Try to show a better preview of sessions and windows in tree mode. 2017-06-30 22:24:08 +00:00
nicm
e028ab3476 Need to flush out the linefeed after wrapper. GitHub issue 970. 2017-06-12 10:57:35 +00:00
nicm
184039044a Typo/style; plus man page escaping from jmc. 2017-06-04 08:02:20 +00:00
nicm
248aa54bfd Style and spacing nits. 2017-05-31 17:56:48 +00:00
nicm
aad4e4ddb1 Rewrite of choose mode, both to simplify and tidy the code and to add
some modern features.

Now the common code is in mode-tree.c, which provides an API used by the
three modes now separated into window-{buffer,client,tree}.c. Buffer
mode shows buffers, client mode clients and tree mode a tree of
sessions, windows and panes.

Each mode has a common set of key bindings plus a few that are specific
to the mode. Other changes are:

- each mode has a preview pane: for buffers this is the buffer content
  (very useful), for others it is a preview of the pane;

- items may be sorted in different ways ('O' key);

- multiple items may be tagged and an operation applied to all of them
  (for example, to delete multiple buffers at once);

- in tree mode a command may be run on the selected item (session,
  window, pane) or on tagged items (key ':');

- displayed items may be filtered in tree mode by using a format (this
  is used to implement find-window) (key 'f');

- the custom format (-F) for the display is no longer available;

- shortcut keys change from 0-9, a-z, A-Z which was always a bit weird
  with keys used for other uses to 0-9, M-a to M-z.

Now that the code is simpler, other improvements will come later.

Primary key bindings for each mode are documented under the commands in
the man page (choose-buffer, choose-client, choose-tree).

Parts written by Thomas Adam.
2017-05-30 21:44:59 +00:00
nicm
e2a18e2b37 Need to store bg for ECH. 2017-05-12 14:56:56 +00:00
nicm
ffd8beb6f6 Need to clear tty context before using it. 2017-05-12 13:29:05 +00:00
nicm
18bab30792 Scrolling needs to use background colour. 2017-05-12 13:00:56 +00:00
nicm
886d50dcab ECH needs to use background colour. 2017-05-12 10:50:11 +00:00
nicm
c0d3f204b0 Clear to start of screen needs to use background colour. 2017-05-11 11:39:30 +00:00
nicm
a2dd7daf4e Fix UTF-8 combining characters in column 0, based on a diff from Keith
Winstein.
2017-04-29 21:27:46 +00:00
nicm
d520dae6ac Make full width panes try to play more nicely with terminal copy and
paste by avoiding explicit line wraps if we think the terminal will wrap
anyway.
2017-04-25 18:30:29 +00:00
nicm
55cd4c7bc7 Can't collect UTF-8 characters of more than one byte at the moment. 2017-04-22 10:30:56 +00:00
nicm
1e6e606f54 Need to flush before writing out cells we are not collecting, also add
some extra logging.
2017-03-07 13:48:28 +00:00
nicm
fba9ebcc0c When redrawing a combined UTF-8 characters in its existing position,
need to save and restore the cursor so that the next character goes into
the right place.
2017-03-06 09:02:36 +00:00
nicm
9e4c5133c8 Scrolling at least needs to be flushed before sending EL to the terminal
(but it is simpler to flush everything, so do that instead).
2017-02-21 10:30:15 +00:00
nicm
c948c6b697 Handle insert cells when cursor at edge of screen correctly, and do a
full flush before insert.
2017-02-16 12:43:08 +00:00
nicm
ddb3750c8b Combining characters need a full flush. 2017-02-09 10:09:14 +00:00
nicm
fd6e46596d When an ordinary (not collected) cell is received, we need to flush any
delayed scrolling before drawing it.
2017-02-09 09:33:15 +00:00
nicm
e100d465da Add support for scroll up escape sequence (CSI S) and use it when
possible instead of sending individual line feeds.
2017-02-08 17:31:09 +00:00
nicm
13a0b6bb3f Collect sequences of printable ASCII characters and process them
together instead of handling them one by one. This is significantly
faster. Sequences are terminated when we reach the end of the line, fill
the internal buffer, or a different character is seen by the input
parser (an escape sequence, or UTF-8).

Rather than writing collected sequences out immediately, hold them until
it is necessary (another screen modification, or we consume all
available data). This means we can discard changes that would have no
effect (for example, lines that would just be scrolled off the screen or
cleared). This reduces the total amount of data we write out to the
terminal - not important for fast terminals, but a big help with slow
(like xterm).
2017-02-08 16:45:18 +00:00
nicm
7475165cd8 Some other tidying bits. 2017-02-08 15:49:29 +00:00
nicm
9cc02d1498 Trying to avoid the occasional newline by saving the last cell on screen
is not actually helping us much and just adds complexity, so don't
bother.
2017-02-08 08:50:10 +00:00
nicm
10e14ae504 Add BCE for clear to start of screen, which was somehow missed. 2017-02-06 19:26:49 +00:00
nicm
3fd34e70e5 Only redraw the modified character when adding combining characters, not
the whole line.
2017-02-06 13:23:00 +00:00
nicm
9e786030df Fix setting the palette of aixterm colours (90-97). 2017-01-12 00:19:32 +00:00
nicm
9a56671a75 Highlight all occurrences of search string after searching in copy mode. 2017-01-05 09:07:15 +00:00
nicm
bee95bf378 Spacing nits. 2016-12-09 21:39:27 +00:00
nicm
8763bced76 Zero dirty count after flushing. 2016-10-18 08:39:18 +00:00
nicm
4179b42424 Add support for BCE (background colour erase). This makes various escape
sequences (notable EL and ED but also IL, DL, ICH, DCH) create blank
cells using the current background colour rather than the default
colour.

On modern systems BCE doesn't really have many benefits, but most other
terminals now support it, some (lazy) applications rely on it, and it is
not hard to include now that we have pane background colours anyway.

Mostly written by Sean Haugh.
2016-10-13 20:27:27 +00:00
nicm
e83ebf50de Fix a couple of problems with insert mode: flush dirty cells before we
modify the screen, not after; and use grid_view_insert_cells to make
space not grid_move_cells.
2016-10-12 15:43:51 +00:00
nicm
4160df4ca4 Redraw selection in tty_draw_line, so it appears when redrawing whole
pane. Reported by Theo Buehler.
2016-10-12 13:24:07 +00:00
nicm
30086e504c screen_write_copy tried to be clever and clear the line if it reached
the end of the source, but it was wrong and causes problems that are
only showing up now we are more aggressive about skipping redundant
screen updates. Remove the optimization entirely as more trouble than it
is worth to fix (and it'll have to go when BCE is done anyway).
2016-10-05 22:00:29 +00:00
nicm
f55c991a1b Wrap some long lines in screen-write.c. 2016-10-05 12:36:36 +00:00
nicm
cf7f3a436a Check padding when writing any character with width > 1, in case they
overlap after the first character (for example with cells xy and ab, y
is replacing a).
2016-09-29 08:50:43 +00:00
nicm
2627ab322e Remember the number of lines scrolled into the history (versus cleared
into the history) and when resizing only use scrolled lines and not
cleared lines (which are probably not intended to reappear). From
Chaoren Lin.
2016-09-02 20:57:20 +00:00
nicm
0f73af876f Don't update cells in each block of data read from a pane immediately,
instead track them as change (dirty) and update them once at the end,
saves much time if repeatedly writing the same cell. Also fix comparison
of cells being equal in a few places (memcmp is not enough).
2016-07-15 00:49:08 +00:00
nicm
aba4438013 Cache selected state so that cells going from selected to unselected are not
skipped, reported by Omar Sandoval.
2016-06-06 07:28:52 +00:00
nicm
36ab4c7c49 Do not draw character to screen if it has not changed, and do not save
last character if it won't be used. This (and last few commits) prompted
by a report from Hubert depesz Lubaczewski.
2016-05-30 09:32:24 +00:00
nicm
344a6a6202 Padding cell is always the same so use a static. 2016-05-27 23:06:12 +00:00
nicm
382222af8e Break the save-last-cell code into a separate function (so it can be
called conditionally later).
2016-05-27 23:02:17 +00:00
nicm
5f2bfd9807 Make the grid_cell passed into screen_write_* const. 2016-04-29 13:21:33 +00:00
nicm
97882f9ce2 Clear RGB flags during selection. 2016-01-31 14:11:49 +00:00
nicm
995af0e2b7 I no longer use my SourceForge address so replace it. 2016-01-19 15:59:12 +00:00
nicm
88bc8f3528 Style nits and line wrapping of function declarations. 2015-12-11 16:37:21 +00:00
nicm
2adf3f42ee Partly revert previous, it is harmless to keep support for UTF-8 mouse
mode inside tmux, just no longer support it for tmux itself.
2015-11-23 23:47:57 +00:00
nicm
32e510bd70 Remove support for the UTF-8 mouse extension. This was a briefly used,
poor idea that was fairly quickly replaced by SGR mouse input (which is
now widespread). It is impossible to tell the difference between UTF-8
and non-UTF-8 mouse input; since the mouse-utf8 option was removed tmux
has not handled it correctly in any case; and it is ridiculous to have
three different forms of mouse input.
2015-11-23 20:53:09 +00:00
nicm
577c0e3e5a Use __unused rather than rolling our own. 2015-11-18 14:27:44 +00:00
nicm
205d15e82d All these return values from utf8_* are confusing, use an enum. 2015-11-14 11:45:43 +00:00
nicm
64333e3ef8 Be more strict about invalid UTF-8. 2015-11-14 10:56:31 +00:00
nicm
c5689a5a40 Long overdue change to the way we store cells in the grid: now, instead
of storing a full grid_cell with UTF-8 data and everything, store a new
type grid_cell_entry. This can either be the cell itself (for ASCII
cells), or an offset into an extended array (per line) for UTF-8
data.

This avoid a large (8 byte) overhead on non-UTF-8 cells (by far the
majority for most users) without the complexity of the shadow array we
had before. Grid memory without any UTF-8 is about half.

The disadvantage that cells can no longer be modified in place and need
to be copied out of the grid and back but it turned out to be lot less
complicated than I expected.
2015-11-13 08:09:28 +00:00
nicm
e71a915412 Rename overly-long utf8data to ud throughout. 2015-11-12 22:04:37 +00:00
nicm
1b86f520ea Nuke the utf8 and status-utf8 options and make tmux only a UTF-8
terminal. We still support non-UTF-8 terminals outside tmux, but inside
it is always UTF-8 (as when the utf8 and status-utf8 options were on).
2015-11-12 11:09:11 +00:00
nicm
16ee4de5df Remove some extra blank lines. 2015-09-14 13:22:02 +00:00
nicm
cc768d77ec Revert to marking lines as wrapped on newlines, fixes problems with
capturep -J.
2015-07-13 13:28:50 +00:00
nicm
8f1302282b Two improvements to reflow from Balazs Kezes:
- Don't extend the line to full width on insert/delete character which
  means leaves extra spaces when reflowing.

- Only mark a line wrapped when the cursor actually goes off the end,
  not on newlines which can be used for positioning.
2014-11-08 12:58:31 +00:00
nicm
0a1a88d63c Better format for printf format attributes. 2014-10-20 23:57:13 +00:00
nicm
2f19df09b1 Copy ACS characters as UTF-8, from Balazs Kezes. 2014-10-02 08:36:26 +00:00
nicm
a5d4b7f3d9 Some more long lines. 2014-04-17 14:45:49 +00:00
nicm
945339b443 Allow replacing each of the many sets of separate foo-{fg,bg,attr}
options with a single foo-style option. For example:

    set -g status-fg yellow
    set -g status-bg red
    set -g status-attr blink

Becomes:

    set -g status-style fg=yellow,bg=red,blink

The -a flag to set can be used to add to rather than replace a style. So:

    set -g status-bg red

Becomes:

    set -ag status-style bg=red

Currently this is fully backwards compatible (all *-{fg,bg,attr} options
remain) but the plan is to deprecate them over time.

From Tiago Cunha.
2014-01-28 23:07:09 +00:00
Nicholas Marriott
772d61f3ed RIS should reset focus reporting, from Hayaki Saito. 2013-05-15 15:32:14 +00:00
Nicholas Marriott
5e4d9a3197 Move the cursor back into the last column on CUU/CUD to match xterm
behaviour. From George Nachman.
2013-03-27 11:19:19 +00:00
Nicholas Marriott
f19836550b Unbreak line wrapping. 2013-03-22 10:41:01 +00:00
Nicholas Marriott
ad5df9bc2f Implement DECAWM (SM/RM 7) using existing MODE_WRAP flag. 2013-03-22 10:36:53 +00:00
Nicholas Marriott
67b4d5b609 Support the latest theory for mouse input, this is enabled/disabled with
SM/RM 1006 and is similar in style to SGR input: \033[<b;x;yM or
\033[b;x;ym. From Egmont Koblinger.
2013-03-22 10:33:50 +00:00
Nicholas Marriott
10682b9e7e Instead of loads of little screen_write_*_on and off functions which
just change mode flags, just have screen_write_mode_set and
screen_write_mode_clear.
2013-03-21 18:47:56 +00:00
Nicholas Marriott
fdbfc7e349 Rather than having two grids for each pane, one for ASCII and one for
UTF-8, collapse the two together. Simplifies the code at the expense of
more memory (which can probably be reduced again later).
2013-01-18 02:16:21 +00:00
Nicholas Marriott
44f8e1caff Implement ECH (erase character, CSI X). Reported by Christian Neukirchen. 2013-01-15 23:18:55 +00:00
Nicholas Marriott
3fa4f691e3 Handle resetting 256-colours properly when parsing #[default],
#[fg=default] and #[bg=default] styles.
2012-12-08 17:05:57 +00:00
Nicholas Marriott
eb0ad181e9 Use ACS characters for choose-tree arrows based on diff from Romain
Francoise.
2012-09-24 12:53:55 +00:00
Nicholas Marriott
df912e3540 xfree is not particularly helpful, remove it. From Thomas Adam. 2012-07-10 11:53:01 +00:00
Nicholas Marriott
0489213b1b Revert screen-write.c r1.54 and fix the bug properly. After wrapping a
line in a pane, the cursor needs to move to the next line unless it
scrolled.
2012-03-17 17:36:03 +00:00
Nicholas Marriott
71df37c247 Store context off before moving the cursor when wrapping, to fix long
standing bug drawing over the status line.
2012-03-14 23:29:07 +00:00
Nicholas Marriott
f4fdddc930 Support "bracketed paste" mode. This adds a -p flag to paste-buffer - if
this is used and the application has requested bracketed pastes, then
tmux surrounds the pasted text by \033[200~ and \033[201~. Applications
like vim can (apparently) use this to avoid, for example, indenting the
text. From Ailin Nemui.
2012-03-03 09:43:22 +00:00
Nicholas Marriott
66f04514cf Add a -R flag to send-keys to reset the terminal. Written ages ago and
Suggested by someone, I forget who.
2012-01-21 08:10:21 +00:00
Nicholas Marriott
f0aad68aee Support for \e[3J to clear the history. Also send the corresponding
terminfo code (E3) before locking.
2011-10-23 10:16:14 +00:00
Nicholas Marriott
96e7f33da3 Support setting the xterm clipboard when copying from copy mode using
the xterm escape sequence for the purpose (if xterm is configured to
allow it).

Written by and much discussed Ailin Nemui, guidance on
xterm/termcap/terminfo from Thomas Dickey.
2011-05-18 20:24:29 +00:00
Nicholas Marriott
044c0f978f Fix character position check, from Tiago Resende. 2011-04-17 19:21:19 +00:00
Nicholas Marriott
d74e5bffba Fix to properly wrap wide characters, from Micah Cowan. 2011-03-26 19:07:33 +00:00
Nicholas Marriott
79e30daeae Support passing through escape sequences to the underlying terminal by
using DCS with a "tmux;" prefix. Escape characters in the sequences must
be doubled. For example:

$ printf '\033Ptmux;\033\033]12;red\007\033\\'

Will pass \033]12;red\007 to the terminal (and change the cursor colour
in xterm). From Kevin Goodsell.
2011-03-07 23:46:27 +00:00
Nicholas Marriott
ecc22c521d When clearing the entire screen, clear lines that are used into the
history like xterm does. Requested ages ago by someone I've forgotten.
2011-01-25 23:40:26 +00:00
Nicholas Marriott
ac3b78a841 Support for UTF-8 mouse input (\033[1005h). This was added in xterm 262
and supports larger terminals than the older way.

If the new mouse-utf8 option is on, UTF-8 mouse input is enabled for all
UTF-8 terminals. The option defaults to on if LANG etc are set in the
same manner as the utf8 option.

With help and based on code from hsim at gmx.li.
2011-01-03 23:35:21 +00:00
Nicholas Marriott
f7c42c21ba Support all four of the xterm mouse modes. Based on a diff from hsim at
gmx.li.
2010-12-29 21:49:06 +00:00
Nicholas Marriott
5afb820d23 Ensure we overwrite UTF-8 wide characters properly, and never overwrite
characters we weren't overlapping.  Fixes "disappearing wide characters"
glitch. From Micah Cowan.
2010-06-21 00:11:12 +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
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
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
a78cc98c8b Cleanup by moving various (mostly horrible) little bits handling UTF-8 grid
data into functions in a new file, grid-utf8.c, and use sizeof intead of
UTF8_DATA.

Also nuke trailing whitespace from tmux.1, reminded by jmc.
2009-11-18 17:02:17 +00:00
Nicholas Marriott
5d7bff4051 A screen can be one cell wide; don't crash if that is the case. 2009-11-16 13:40:45 +00:00
Nicholas Marriott
9df4e7597d Add an explicit zero-length check for UTF-8 input data, prompted by a report
from parfait via deraadt.

While here, add a statement to set the width when filling with _s if not enough
space (width should never be high enough at the moment anyway), and wrap some
long lines.
2009-11-11 18:53:21 +00:00
Nicholas Marriott
1af09d6330 Try to reduce the UTF-8 mess.
Get rid of passing around u_char[4]s and define a struct utf8_data which has
character data, size (sequence length) and width. Move UTF-8 character
collection into two functions utf8_open/utf8_append in utf8.c which fill in
this struct and use these functions from input.c and the various functions in
screen-write.c.

Space for rather more data than is necessary for one UTF-8 sequence is in the
utf8_data struct because screen_write_copy is still nasty and needs to reinject
the character (after combining) into screen_write_cell.
2009-10-20 19:18:28 +00:00
Nicholas Marriott
62f234ce3b UTF-8 combined character fixes.
Thai can have treble combinations (1 x width=1 then 2 x width=0) so bump the
UTF-8 cell data size to 9 and alter the code to allow this.

Also break off the combining code into a separate function, handle any further
combining beyond the buffer size by replacing the character with _s, and when
redrawing the UTF-8 character don't assume the first part has just been
printed, redraw the entire line.
2009-10-20 17:33:33 +00:00
Nicholas Marriott
387f4d42cc Move the check for whether to force a line wrapper lower down into the tty code
where it has access to the tty width, which is what should have been checked.
2009-10-20 16:32:23 +00:00
Nicholas Marriott
daa26079ee Always move the cursor position on !xenl terminals, since there is no invisible
last cursor position.

Also nuke an unused variable.
2009-10-17 08:35:38 +00:00
Nicholas Marriott
fe26b5d25f Don't print wide characters at screen width - 1. Matches uterm behaviour and
is probably a better idea anyway.
2009-10-17 08:32:18 +00:00
Nicholas Marriott
43d62c1ae3 Instead of having a complicated check to see if the cursor is in the last
position to avoid an explicit wrap, actually move it there.

Some UTF-8 fixes to come.
2009-10-17 08:24:46 +00:00