Update to 6.0:
This release is designed to be source-compatible with ncurses 5.0
through 5.9; providing a new application binary interface (ABI).
Although the source can still be configured to support the ncurses 5
ABI, the intent of the release is to provide extensions which are
generally useful, but binary-incompatible with ncurses 5:
* Extend the cchar_t structure to allow more than 16 colors to be
* Modify the encoding of mouse state to make room for a 5th mouse
button. That allows one to use ncurses with a wheel mouse with
xterm or similar X terminal emulators.
There are, of course, numerous other improvements, including
* fixes made based on the Clang and Coverity static analyzers.
* memory leak fixes using Valgrind
The release notes mention some bug-fixes, but are focused on new
features and improvements to existing features log since ncurses 5.9
While the intent of the release is to provide a new stable ABI, there
are other development activities which are summarized below.
* The original release plan, e.g., for "5.10" was to improve the
MinGW port. Ultimately that was completed (wide-character support,
mouse, etc), but was deferred to focus on termcap support and
performance issues. Also, pinpointing the problems with Console2
took a while.
* A review of termcap compatibility in 2011 led to several minor
fixes in the library and improvements to utilities. To do this
properly, a review of the various extent termcap implementations
The termcap library checker (tctest) (not part of ncurses) was
one result. A followup review of performance using
ncurses-examples in 2014 led to additional improvements.
* Output buffering provided a further, but worthwhile distraction. A
bug report in 2012 regarding the use of signal handlers in
ncurses) pointed out a problem with the use of unsafe
functions for handling SIGTSTP. Other signals could be addressed
with workarounds; repairing SIGTSTP required a different approach.
The solution required changing internal behavior of the library:
how it handles output buffering.
Now ncurses buffers its own output, independently of the standard
output. A few applications relied upon the library's direct reuse
of the standard output buffering; however that is unspecified
behavior and has never been a recommended practice. Identifying
these applications as well as refining the change to permit
low-level applications to work consistently took time.
* Since the introduction of the experimental support for 256 colors
early in 2005 (released in ncurses 5.5), there has been
increased user interest. Almost all packagers continue providing
the ncurses ABI 5 which cannot support 256 colors.
* Symbol versioning, or the lack of it in ncurses, is the main
reason why packagers would be reluctant to add a new ncurses ABI.
This release provides the new ABI along with script-generated
lists of versioned symbols which can be used for both ABI 5 and 6
(with distinct names to keep the two separate). This took time to
development, as reported in Symbol versioning in ncurses.
X/Open curses provides more than one initialization function:
* initscr (the simplest) accepts no parameters.
* newterm accepts parameters for the stream input and output
* setupterm (the low-level function) accepts a parameter for the
file descriptor of the output.
They are documented in X/Open as if initscr calls newterm using stdout
for output stream, and in turn newterm calls setupterm using
fileno(stdout) for the file descriptor. As long as an implementation
acts as if it does this, it conforms. In practice, implementations do
what is implied. This creates a problem: the low-level setupterm
function's file descriptor is unbuffered, while newterm implies
buffered output. X/Open Curses says that all output is done through
the file descriptor, and does not say how the output stream is
Initially, ncurses used the file descriptor (obtained from the output
stream passed to newterm) for changing the terminal modes, and relied
upon the output parameter of newterm for buffered output. Later (to
avoid using unsafe buffered I/O in signal handlers), ncurses was
modified to use the file descriptor (unbuffered output) when cleaning
up on receipt of a signal. Otherwise (when not handling a signal), it
continued to use the buffered output.
That approach worked reasonably well and as a side effect, using the
same buffered output as an application might use for printf meant that
no flushing was needed when switching between normal- and
There were a couple of problems:
* to get good performance, curses (not only ncurses, but SVr4 curses
in general) set an output buffer using setbuf or similar function.
There is no standard (or portable) way to turn that output buffer
off, and revert to line-buffering. The NCURSES_NO_SETBUF
environment variable did make it optional.
* to handle SIGTSTP (the "stop" signal), ncurses relied upon unsafe
functions. That is, due to the complexity of the feature, it
relied upon reusing existing functions which should not have been
called via the signal handler.
Conveniently, solving the second problem (by making ncurses do its own
output buffering) also fixed the first one. But there were special
cases to resolve: low-level functions such as mvcur, putp, vidattr
explicitly use the standard output. Those functions were reused
internally, and required modification to distinguish whether they were
used by the high-level or low-level interfaces.
Finally, there may still be a few programs which should be modified to
improve their portability, e.g., adding an
when switching from "shell" mode to "program" \
Those are fairly rare because most programmers have learned not to mix
printf and printw.
This release introduces symbol-versioning to ncurses because without
it, the change of ABI would be less successful. A lengthy discussion
will be presented in Symbol versioning in ncurses. These notes
summarize what has changed, and what can be done with the new release.
Symbol-versioning allows the developers of a library to mark each
public symbol (both data and functions) with an identifier denoting
the library name and the version for which it was built. By doing
this, users of the library have a way to help ensure that applications
do not accidentally load an incompatible library. In addition, private
symbols can be hidden entirely.
This release provides sample files for the four principal
configurations of ncurses libraries: ncurses, ncursesw, ncursest and
ncursestw. Each sample is given in two forms:
These list all public symbols, together with version names.
These list all public symbols, without version names.
The sample files are generated by scripts which take into account a
few special cases such as tack to omit many of the ncurses private
symbols (beginning with "_nc_"). Here are counts of globals versus
Config Symbols Globals Locals "_nc_"
ncurses 976 796 180 332
ncursesw 1089 905 184 343
ncursest 979 804 175 358
ncursestw 1098 914 184 372
Although only four sample configurations are presented, each is formed
by merging symbols from several combinations of configure-script
options, taking into account advice from downstream packagers. Because
they are formed by merging, the sample files may list a symbol which
is not in a given package. That is expected. The samples have been
tested and are working with systems (such as Fedora, FreeBSD and
Debian) which fully support this feature. There are other systems
which do not support the feature, and a few (such as Solaris) which
provide incomplete support.
The version-naming convention used allows these sample files to build
distinct libraries for ABI 5 and 6. Version names consist of
* configuration name, e.g., "NCURSESW" for the wide-character
* ABI version (if not 5)
* library name for two special cases which have the same interface
across configurations: "TINFO" and "TIC"
* release version
* patch date (for the release version)
For example, running nm -D on the libraries in the ncurses6 test
package shows these symbol-versions:
0000000000000000 A NCURSES6_TIC_5.0.19991023
0000000000000000 A NCURSES6_TIC_5.1.20000708
0000000000000000 A NCURSES6_TIC_5.5.20051010
0000000000000000 A NCURSES6_TIC_5.7.20081102
0000000000000000 A NCURSES6_TIC_5.9.20150530
0000000000000000 A NCURSES6_TINFO_5.0.19991023
0000000000000000 A NCURSES6_TINFO_5.1.20000708
0000000000000000 A NCURSES6_TINFO_5.2.20001021
0000000000000000 A NCURSES6_TINFO_5.3.20021019
0000000000000000 A NCURSES6_TINFO_5.4.20040208
0000000000000000 A NCURSES6_TINFO_5.5.20051010
0000000000000000 A NCURSES6_TINFO_5.6.20061217
0000000000000000 A NCURSES6_TINFO_5.7.20081102
0000000000000000 A NCURSES6_TINFO_5.8.20110226
0000000000000000 A NCURSES6_TINFO_5.9.20150530
0000000000000000 A NCURSESW6_5.1.20000708
0000000000000000 A NCURSESW6_5.3.20021019
0000000000000000 A NCURSESW6_5.4.20040208
0000000000000000 A NCURSESW6_5.5.20051010
0000000000000000 A NCURSESW6_5.6.20061217
0000000000000000 A NCURSESW6_5.7.20081102
0000000000000000 A NCURSESW6_5.8.20110226
0000000000000000 A NCURSESW6_5.9.20150530
As a special case, this release (which makes the final change for ABI
5) is marked with release version 5.9 and patch date 20150530.
The new release has several improvements for performance and building.
* several files in ncurses- and progs-directories were modified to
allow const data used in internal tables to be put by the linker
into the readonly text segment.
* various improvements were made to building the Ada95 binding, both
in simplifying the generated files as well as improving the way it
There are also new features in the libraries:
* added use_tioctl function
* added wgetdelay to retrieve _delay member of WINDOW if it
happens to be opaque, e.g., in the pthread configuration.
* added A_ITALIC extension.
* added form library extension O_DYNAMIC_JUSTIFY option which
can be used to override the different treatment of justification
for static versus dynamic fields .
* rewrote putwin and getwin, making an extended version
which is capable of reading screen-dumps between the wide/normal
ncurses configurations. These are text files, except for a magic
code at the beginning:
0 string \210\210 Screen-dump (ncurses)
* several changes to mouse support include:
+ added decoder for xterm SGR 1006 mouse mode.
+ added experimental support for "%u" format to terminfo.
+ improved behavior of wheel-mice for xterm protocol: noting
that there are only button-presses for buttons "4" and \
so there is no need to wait to combine events into
There are a few new configure options dealing with library
* add "--enable-ext-putwin" configure option to turn on the extended
putwin/getwin. By default, this is enabled for ABI 6 and disabled
with ABI 5.
* add "--enable-string-hacks" option to control whether strlcat and
strlcpy may be used. Because ncurses already does the requisite
buffer-limit checks, this feature is mainly of interest to quiet
compiler-warnings on a few systems.
* add configure option "--with-tparm-arg" to allow tparm's
parameters to be something more likely to be the same size as a
pointer, e.g., intptr_t (again, the default is set for ABI 6).
Most of the termcap-related changes based on development of tctest
(termcap library checker) are implemented in the tic and infocmp
programs rather than affecting the library. As noted in the
discussion of tctest, ncurses's ability to translate between
terminfo and termcap formats has been improved at different times, but
subject to feedback from "real" termcap users. There are very few of
those. Nowadays, virtually all termcap users are using ncurses (or
NetBSD, with its own terminfo library) and their programs are actually
using terminfo rather than termcap data.
Still, there are a few. A comment about the translation of the ASCII
NUL character prompted a review:
* Both terminfo and termcap store string capabilities as
* In terminfo, a \0 in a terminal description is stored as \200.
* There are no (known) terminals which would behave differently when
sent \0 or \200.
* When translating to terminfo format (or displaying a printable
version of an entry using infocmp), ncurses shows \200 as \0.
* It has done this since 1998 (quoting from the NEWS file):
+ modify _nc_tic_expand() to generate \0 rather than \200.
+ correct translation of terminfo "^@", to \200, like \0.
* However, the _nc_tic_expand function (which optionally produces
terminfo or termcap format) did not address this special case for
termcap. Even the later 4.4BSD cgetstr interprets a \0
literally, ending that string (rather than using the terminfo
As a result of the review, several improvements were made to ncurses
translation to/from termcap format -- and improving the checks made in
tic for consistency of entries. Most of these are not of general
interest, except for two new command-line options for tic and infocmp:
* the "-0" option generates termcap/terminfo source on a single
* the "-K" option provides stricter BSD-compatibility for termcap
Other user-visible improvements and new features include:
* added "-D" option to tic and infocmp, to show the database
locations that it could use.
* added "-s" option to toe, to sort its output.
* extended "-c" and "-n" options of infocmp to allow \
than two entries.
* modified toe's report when "-a" and "-s" options are \
add a column showing which entries belong to a given database.
* modified the clear program to take into account the "E3" extended
capability to clear the terminal's scrollback buffer.
Along with the library and utilities, many improvements were made to
the ncurses-examples. Some were made to allow building (and
comparison-testing) against NetBSD curses and PDCurses. Both lack some
of the X/Open Curses features, necessitating customization. But this
activity was useful because it showed some remaining performance
issues (which have been resolved in this release).
These changes were made to verify compatibility or compare performance
* made workarounds for compiling test-programs with NetBSD curses,
though it lacks some common functions such as use_env.
* added dots_termcap test-program
* added dots_curses test-program, for comparison with the low-level
* added test_setupterm test-proram to demonstrate normal/error
returns from the setupterm and restartterm functions.
* added "-d", "-e" and "-q" options to \
the demo_terminfo and
* added "-y" option to demo_termcap and test/demo_terminfo
test-programs to demonstrate behavior with/without extended
* modified demo_termcap and demo_terminfo test-programs to make
their options more directly comparable, and add "-i" option to
specify a terminal description filename to parse for names to
* rewrote the tests for mvderwin and test for recursive
mvwin in the movewindow test-program.
These changes were made to help with the MinGW port:
* added test-screens to the ncurses test-program to show
256-characters at a time, to help with MinGW port.
* modified the view test-program to load UTF-8 when built with MinGW
by using regular win32 API because the MinGW functions mblen and
mbtowc do not work.
* added "-s" option to the view test-program to allow it to start in
single-step mode, reducing size of trace files when it is used for
debugging MinGW changes.
These changes were made to verify new extensions in ncurses:
* added form_driver_w entrypoint to wide-character forms
library, as well as form_driver_w test-program.
* modified ncurses test-program's b/B tests to display lines only
for the attributes which a given terminal supports, to make room
for an italics test.
* modified ncurses test-program, adding "-E" and \
"-T" options to
demonstrate use_env versus use_tioctl.
* modified ncurses test-program's c/C tests to cycle through subsets
of the total number of colors, to better illustrate
8/16/88/256-colors by providing directly comparable screens.
* modified the ncurses test-program to also show position reports in
These changes were made to make the examples more useful:
* added scripts for building dpkg and rpm test-packages
* modified the hanoi test-program to show the minimum number of
moves possible for the given number of tiles.
* modified the knight test-program to show the number of choices
possible for each position in automove option, e.g., to allow user
to follow Warnsdorff's rule to solve the puzzle.
This release provides improvements to tic's "-c" checking option,
which was used for example to
* make sgr in several entries agree with other caps.
* correct padding in some entries where earlier versions had
miscounted the number of octal digits.
There are several new terminal descriptions:
* mlterm is now aliased to mlterm3
* nsterm is now derived from nsterm-256color
* teken is FreeBSD's "xterm" console.
* tmux is derived from screen.
* several screen.XXX entries support the respective variations for
* simpleterm is now 0.5
* vte is aliased to vte-2012
A few entries use extensions (user-defined terminal capabilities):
* E3, used in linux, putty and xterm-basic is tested in the
clear program to erase a terminal's scrollback.
* TS is used in the xterm+sl building block to help deprecate
the misuse of tsl for xterm's title-string.
* XT is used in some terminfo entries to improve usefulness for
other applications than screen, which would like to pretend that
xterm's title is a status-line.
* xm is used in examples xterm-1005 and xterm-1006 to
illustrate a way to make mouse handling more general
A few terminals support italics and/or dim capabilities. In
particular, screen does not. Documented that, and accommodated the
terminals where this feature works with the A_ITALIC extension.
* konsole, mlterm3 (italics)
* nsterm (dim)
* screen (dim)
* vte (dim, italics)
* xterm (dim, italics)
As usual, this release
* improves documentation by describing new features,
* attempts to improve the description of features which users have
* fills in overlooked descriptions of features which were described
in the NEWS file but treated sketchily in manual pages.
In addition, the mechanism for producing HTML versions of the
documentation has been improved:
* use an improved version of man2html to generate html manpages.
* regenerated NCURSES-Programming-HOWTO.html to fix some of the
broken html emitted by docbook.
* Ada95 binding:
+ modify makefile rules to ensure that the PIC option is not
used when building a static library
+ make Ada95 build-fix for big-endian architectures such as
sparc. This undoes one of the fixes from 20110319, which
added an "Unused" member to representation clauses, replacing
that with pragmas to suppress warnings about unused bits.
* Color and attributes:
+ parenthesize parameter of COLOR_PAIR and PAIR_NUMBER in
curses.h in case it happens to be a comma-expression.
+ improve 20021221 workaround for broken acs, handling a
case where that ACS_xxx character is not in the acsc string
but there is a known wide-character which can be used.
+ modify init_pair to accept -1's for color value after
assume_default_colors has been called.
+ add a check in start_color to limit color-pairs to 256
when extended colors are not supported.
* Resizing the screen:
+ propagate error-returns from wresize, i.e., the internal
increase_size and decrease_size functions through
+ add check for zero/negative dimensions for resizeterm and
+ modify resizeterm to always push a KEY_RESIZE onto the fifo,
even if screensize is unchanged. Modify library to push a
KEY_RESIZE if there was a SIGWINCH, even if it does not call
resizeterm). These changes eliminate the case where a
SIGWINCH is received, but ERR is returned from wgetch or
wgetnstr because the screen dimensions did not change.
* Low-level interfaces
+ fix an old bug in the termcap emulation; "%i" was ignored in
tparm because the parameters to be incremented were already
on the internal stack.
+ change "%l" behavior in tparm to push the string length onto
the stack rather than saving the formatted length into the
+ modify name-comparison for tgetstr, etc., to accommodate
legacy applications as well as to improve compatbility with
BSD 4.2 termcap implementations (see note for 980725).
* High-level interfaces
+ modify internal recursion in wgetch which handles cooked mode
to check if the call to wgetnstr returned an error. This can
happen when both nocbreak and nodelay are set, for instance
(see note for 960418).
+ add a check in internal function waddch_nosync to ensure that
tab characters are treated as control characters; some broken
locales claim they are printable.
+ modify menu library to ensure that a menu's top-row is
adjusted as needed to ensure that the current item is on the
+ fix special case where double-width character overwrites a
single- width character in the first column.
The ncurses 6.0 configure script makes changes to the default value of
several configure options, depending on the --with-abi-version option
(i.e., whether its value is "5" or "6"):
Feature introduced in 970405 supports the use of const
where X/Open Curses should have, but did not. NetBSD curses
does something similar with const.
Extends the cchar_t structure to allow more than 16 colors to
be encoded. This applies only to the wide-character
Modifies the encoding of mouse state to make room for a 5th
mouse button. That allows one to use ncurses with a wheel mouse
with xterm or similar X terminal emulators.
Modifies the file-format written by putwin to use printable
text rather than binary files, allowing getwin to read screen
dumps written by differently-configured ncurses libraries. The
extended getwin can still read binary screen dumps from the
same configuration of ncurses. This does not change the ABI
(the binary interface seen by calling applications).
Modifies the FIELDTYPE structure used for the form library to
make it more generic.
Allows an application to define _LP64 to declare chtype and
mmask_t as simply "unsigned" rather than the configured types
using the --with-chtype and --with-mmask_t options.
Compile-in support for extended functions which accept a SCREEN
pointer, reducing the need for juggling the global SP value
with set_term and delscreen.
Makes chtype explicitly a 32-bit unsigned value.
Makes mmask_t explicitly a 32-bit unsigned value.
X/Open Curses declares tparm using long for each of the
parameters aside from the formatting string, presuming that
long and char* are the same size. This configure option uses
intptr_t which provides a better guarantee of the sizes.
The configure script no longer checks for antique compilers; c89 is
assumed as a minimum. There are a few features from later revisions
which are used when available. The configure script makes checks to
turn on useful warnings from clang, gcc and icc. You should be able to
build ncurses 6.0 with any of the current (or not so current) C
compilers available in 2015.
The configure script, by the way, makes changes which do not work with
systems whose /bin/sh is non-POSIX. This mainly affects Solaris (the
other vendor unix systems have followed the POSIX guidelines for the
past twenty years). If you must build on Solaris, its xpg4
binaries suffice, e.g.,
WHAT=`hostname|sed -e 's/\..*//'`
cat >>$OUT <<EOF/
** node: $WHAT
** user: `id`
** conf: $*
if test -f /usr/xpg4/bin/sh
rm -f config.status config.cache
$SHELL ./configure --verbose \
--prefix=$TOP $* 2>&1 | tee -a $OUT
Other major changes to the configure script include:
* ABI 6 is now the default, intending that the existing ABI 5 should
build as before using the "--with-abi-version=5" option.
* added --with-extra-suffix option to help with installing
nonconflicting ncurses6 packages, e.g., avoiding header- and
NOTE: as a side-effect, this renames
adacurses-config to adacurses5-config and
adacursesw-config to adacursesw5-config
* the configure script looks for gnatgcc if the Ada95 binding is
built, in preference to the default gcc/cc. The script also
ensures that the Ada95 binding is built with the level of
optimization as the C libraries.
* the configure script captures define's related to -D_XOPEN_SOURCE
from the configure check and adds those to the *-config and *.pc
files, to simplify use for the wide-character libraries.