Background
Before ncurses 5.0, terminfo databases used a fixed repertoire of
terminal capabilities designed for the SVr2 terminal database in
1984, and extended in stages through SVr4 (1989), and
standardized in the Single Unix Specification beginning in 1995.
Most of the extensions in this fixed repertoire were additions to
the tables of boolean, numeric and string capabilities. Rather
than change the meaning of an existing capability, a new name was
added. The terminfo database uses a binary format; binary
compatibility was ensured by using a header which gave the number
of items in the tables for each type of capability. The
standardization was incomplete:
• The binary format itself is not described in the X/Open
Curses documentation. Only the source format is described.
Library developers rely upon the SVr4 documentation, and
reverse-engineering the compiled terminfo files to match the
binary format.
• Lacking a standard for the binary format, most
implementations copy the SVr2 binary format, which uses
16-bit signed integers, and is limited to 4096-byte entries.
The format cannot represent very large numeric capabilities,
nor can it represent large numbers of special keyboard
definitions.
• The tables of capability names differ between
implementations.
Although they may provide all of the standard capability
names, the position in the tables differs because some
features were added as needed, while others were added (out
of order) to comply with X/Open Curses.
While ncurses' repertoire of predefined capabilities is
closest to Solaris, Solaris's terminfo database has a few
differences from the list published by X/Open Curses. For
example, ncurses can be configured with tables which match
the terminal databases for AIX, HP-UX or OSF/1, rather than
the default Solaris-like configuration.
• In SVr4 curses and ncurses, the terminal database is defined
at compile-time using a text file which lists the different
terminal capabilities.
In principle, the text-file can be extended, but doing this
requires recompiling and reinstalling the library. The text-
file used in ncurses for terminal capabilities includes
details for various systems past the documented X/Open Curses
features. For example, ncurses supports these capabilities
in each configuration:
memory_lock
(meml) lock memory above cursor
memory_unlock
(memu) unlock memory
box_chars_1
(box1) box characters primary set
The memory lock/unlock capabilities were included because
they were used in the X11R6 terminal description for xterm.
The box1 capability is used in @TIC@ to help with terminal
descriptions written for AIX.
During the 1990s, some users were reluctant to use terminfo in
spite of its performance advantages over termcap:
• The fixed repertoire prevented users from adding features for
unanticipated terminal improvements (or required them to
reuse existing capabilities as a workaround).
• The limitation to 16-bit signed integers was also mentioned.
Because termcap stores everything as a string, it could
represent larger numbers.
Although termcap's extensibility was rarely used (it was never
the speaker who had actually used the feature), the criticism had
a point. ncurses 5.0 provided a way to detect nonstandard
capabilities, determine their type and optionally store and
retrieve them in a way which did not interfere with other
applications. These are referred to as user-defined capabilities
because no modifications to the toolset's predefined capability
names are needed.
The ncurses utilities @TIC@ and @INFOCMP@ have a command-line
option '-x' to control whether the nonstandard capabilities are
stored or retrieved. A library function use_extended_names is
provided for the same purpose.
When compiling a terminal database, if '-x' is set, @TIC@ will
store a user-defined capability if the capability name is not one
of the predefined names.
Because ncurses provides a termcap library interface, these user-
defined capabilities may be visible to termcap applications:
• The termcap interface (like all implementations of termcap)
requires that the capability names are 2-characters.
When the capability is simple enough for use in a termcap
application, it is provided as a 2-character name.
• There are other user-defined capabilities which refer to
features not usable in termcap, e.g., parameterized strings
that use more than two parameters or use more than the
trivial expression support provided by termcap. For these,
the terminfo database should have only capability names with
3 or more characters.
• Some terminals can send distinct strings for special keys
(cursor-, keypad- or function-keys) depending on modifier
keys (shift, control, etc.). While terminfo and termcap have
a set of 60 predefined function-key names, to which a series
of keys can be assigned, that is insufficient for more than a
dozen keys multiplied by more than a couple of modifier
combinations. The ncurses database uses a convention based
on xterm to provide extended special-key names.
Fitting that into termcap's limitation of 2-character names
would be pointless. These extended keys are available only
with terminfo.
Recognized capabilities
The ncurses library uses the user-definable capabilities. While
the terminfo database may have other extensions, ncurses makes
explicit checks for these:
AX boolean, asserts that the terminal interprets SGR 39 and
SGR 49 by resetting the foreground and background color,
respectively, to the default.
This is a feature recognized by the screen program as well.
E3 string, tells how to clear the terminal's scrollback
buffer. When present, the clear(1) program sends this
before clearing the terminal.
The command 'tput clear' does the same thing.
RGB
boolean, number or string, to assert that the
set_a_foreground and set_a_background capabilities
correspond to direct colors, using an RGB (red/green/blue)
convention. This capability allows the color_content
function to return appropriate values without requiring the
application to initialize colors using init_color.
The capability type determines the values which ncurses
sees:
boolean
implies that the number of bits for red, green and blue
are the same. Using the maximum number of colors,
ncurses adds two, divides that sum by three, and assigns
the result to red, green and blue in that order.
If the number of bits needed for the number of colors is
not a multiple of three, the blue (and green) components
lose in comparison to red.
number
tells ncurses what result to add to red, green and blue.
If ncurses runs out of bits, blue (and green) lose just
as in the boolean case.
string
explicitly list the number of bits used for red, green
and blue components as a slash-separated list of decimal
integers.
Because there are several RGB encodings in use,
applications which make assumptions about the number of
bits per color are unlikely to work reliably. As a trivial
case, for example, one could define RGB#1 to represent the
standard eight ANSI colors, i.e., one bit per color.
U8 number, asserts that ncurses must use Unicode values for
line-drawing characters, and that it should ignore the
alternate character set capabilities when the locale uses
UTF-8 encoding. For more information, see the discussion
of NCURSES_NO_UTF8_ACS in ncurses(3X).
Set this capability to a nonzero value to enable it.
XM string, override ncurses's built-in string which
enables/disables xterm mouse mode.
ncurses sends a character sequence to the terminal to
initialize mouse mode, and when the user clicks the mouse
buttons or (in certain modes) moves the mouse, handles the
characters sent back by the terminal to tell it what was
done with the mouse.
The mouse protocol is enabled when the mask passed in the
mousemask function is nonzero. By default, ncurses handles
the responses for the X11 xterm mouse protocol. It also
knows about the SGR 1006 xterm mouse protocol, but must to
be told to look for this specifically. It will not be able
to guess which mode is used, because the responses are
enough alike that only confusion would result.
The XM capability has a single parameter. If nonzero, the
mouse protocol should be enabled. If zero, the mouse
protocol should be disabled. ncurses inspects this
capability if it is present, to see whether the 1006
protocol is used. If so, it expects the responses to use
the SGR 1006 xterm mouse protocol.
The xterm mouse protocol is used by other terminal
emulators. The terminal database uses building-blocks for
the various xterm mouse protocols which can be used in
customized terminal descriptions.
The terminal database building blocks for this mouse
feature also have an experimental capability xm. The 'xm'
capability describes the mouse response. Currently there
is no interpreter which would use this information to make
the mouse support completely data-driven.
xm shows the format of the mouse responses. In this
experimental capability, the parameters are
p1 y-ordinate
p2 x-ordinate
p3 button
p4 state, e.g., pressed or released
p5 y-ordinate starting region
p6 x-ordinate starting region
p7 y-ordinate ending region
p8 x-ordinate ending region
Here are examples from the terminal database for the most
commonly used xterm mouse protocols:
xterm+x11mouse|X11 xterm mouse protocol,
kmous=\E[M, XM=\E[?1000%?%p1%{1}%=%th%el%;,
xm=\E[M
%?%p4%t%p3%e%{3}%;%' '%+%c
%p2%'!'%+%c
%p1%'!'%+%c,
xterm+sm+1006|xterm SGR-mouse,
kmous=\E[<, XM=\E[?1006;1000%?%p1%{1}%=%th%el%;,
xm=\E[<%i%p3%d;
%p1%d;
%p2%d;
%?%p4%tM%em%;,
Extended key-definitions
Several terminals provide the ability to send distinct strings
for combinations of modified special keys. There is no standard
for what those keys can send.
Since 1999, xterm has supported shift, control, alt, and meta
modifiers which produce distinct special-key strings. In a
terminal description, ncurses has no special knowledge of the
modifiers used. Applications can use the naming convention
established for xterm to find these special keys in the terminal
description.
Starting with the curses convention that key names begin with 'k'
and that shifted special keys are an uppercase name, ncurses'
terminal database defines these names to which a suffix is added:
Name Description
───────────────────────────────────────────────────────────────
kDC special form of kdch1 (delete character)
kDN special form of kcud1 (cursor down)
kEND special form of kend (End)
kHOM special form of khome (Home)
kLFT special form of kcub1 (cursor-left or cursor-back)
kNXT special form of knext (Next, or Page-Down)
kPRV special form of kprev (Prev, or Page-Up)
kRIT special form of kcuf1 (cursor-right, or cursor-forward)
kUP special form of kcuu1 (cursor-up)
These are the suffixes used to denote the modifiers:
Value Description
──────────────────────────────────
2 Shift
3 Alt
4 Shift + Alt
5 Control
6 Shift + Control
7 Alt + Control
8 Shift + Alt + Control
9 Meta
10 Meta + Shift
11 Meta + Alt
12 Meta + Alt + Shift
13 Meta + Ctrl
14 Meta + Ctrl + Shift
15 Meta + Ctrl + Alt
16 Meta + Ctrl + Alt + Shift
None of these are predefined; terminal descriptions can refer to
names which ncurses will allocate at runtime to key-codes. To
use these keys in an ncurses program, an application could do
this:
• using a list of extended key names, ask tigetstr(3X) for
their values, and
• given the list of values, ask key_defined(3X) for the key-
code which would be returned for those keys by wgetch(3X).
