NCurses: Why does init_color return OK but still not set the color?

Question

I am attempting to set the background of a block of text to red by setting the value COLOR_BLUE to the rgb values of: (1000, 0, 0). The background is correct (RED) if I use Gnome's "Terminal" application, but incorrect (BLUE) if I use KDE's "Konsole". Why is this? The following code will compile on Linux with the compile line:

g++ filename.cpp -lcurses

I set my TERM type to xterm-256color via:

export TERM=xterm-256color

here's the code:

#include <ncurses.h>
#include <cassert>
#include <csignal>

static bool stop = false;

void sigAbortHandler(int _sig)
{
    stop = true;
}

int main(int _argc, char **_argv)
{
    signal(SIGABRT, &sigAbortHandler);

    WINDOW *window = initscr();

    if (!has_colors())
    {
        delwin(window);
        endwin();
        perror("You must enable colors in your console");
    }

    if (!can_change_color())
    {
        delwin(window);
        endwin();
        perror("Error: unable to change colors, "
               "trying setting your TERM type to enable colors");
    }

    assert(start_color() == OK);
    keypad(stdscr, TRUE);
    cbreak();
    noecho();
    curs_set(0);
    nodelay(window, true);

    int background = COLOR_BLUE;
    assert(init_color(background, 1000, 0, 0) == OK);
    int foreground = 2;
    assert(init_color(foreground, 0, 0, 0) == OK);
    int pair = 1;
    assert(init_pair(pair, foreground, background) == OK);
    assert(wattron(window, COLOR_PAIR(pair)) == OK);

    short r, g, b;
    color_content(background, &r, &g, &b);
    assert(mvwprintw(window, 10, 10, "color content: %d, %d, %d", r, g, b) == OK);
    assert(wrefresh(window) == OK);
    assert(wattroff(window, COLOR_PAIR(pair)) == OK);

    while (!stop)
    {
    }

    delwin(window);
    endwin();

    return 0;
}

Answer 1

The problem is that Konsole does not implement the feature you are expecting. ncurses only knows what is in the terminal description which you have configured. There is no general-purpose method for checking that it is correct. The RETURN-VALUE section of init_color says

init_color
returns an error if the terminal does not support this feature, e.g., if the initialize_color capability is absent from the terminal description.

Presumably you have something like TERM=xterm-256color. The ncurses terminal database has a correct konsole-256color, which (using infocmp to compare with xterm-256color) differs in many respects:

comparing xterm-256color to konsole-256color.
    comparing booleans.
    ccc: T:F.
    km: T:F.
    mc5i: T:F.
    comparing numbers.
    comparing strings.
    bel: '^G', NULL.
    cbt: '\E[Z', NULL.
    cnorm: '\E[?12l\E[?25h', '\E[?25h'.
    cvvis: '\E[?12;25h', NULL.
    dim: '\E[2m', NULL.
    el1: '\E[1K', NULL.
    enacs: NULL, '\E)0'.
    ich: '\E[%p1%d@', NULL.
    initc: '\E]4;%p1%d;rgb\:%p2%{255}%*%{1000}%/%2.2X/%p3%{255}%*%{1000}%/%2.2X/%p4%{255}%*%{1000}%/%2.2X\E\\', NULL.
    invis: '\E[8m', NULL.
    is2: '\E[!p\E[?3;4l\E[4l\E>', '\E[m\E[?7h\E[4l\E>\E7\E[r\E[?1;3;4;6l\E8'.
    kDC: '\E[3;2~', NULL.
    kEND: '\E[1;2F', NULL.
    kHOM: '\E[1;2H', NULL.
    kIC: '\E[2;2~', NULL.
    kLFT: '\E[1;2D', NULL.
    kNXT: '\E[6;2~', NULL.
    kPRV: '\E[5;2~', NULL.
    kRIT: '\E[1;2C', NULL.
    kb2: '\EOE', NULL.
    kent: '\EOM', NULL.
    kf13: '\E[1;2P', '\EO2P'.
    kf14: '\E[1;2Q', '\EO2Q'.
    kf15: '\E[1;2R', '\EO2R'.
    kf16: '\E[1;2S', '\EO2S'.
    kf25: '\E[1;5P', '\EO5P'.
    kf26: '\E[1;5Q', '\EO5Q'.
    kf27: '\E[1;5R', '\EO5R'.
    kf28: '\E[1;5S', '\EO5S'.
    kf37: '\E[1;6P', '\EO6P'.
    kf38: '\E[1;6Q', '\EO6Q'.
    kf39: '\E[1;6R', '\EO6R'.
    kf40: '\E[1;6S', '\EO6S'.
    kf49: '\E[1;3P', '\EO3P'.
    kf50: '\E[1;3Q', '\EO3Q'.
    kf51: '\E[1;3R', '\EO3R'.
    kf52: '\E[1;3S', '\EO3S'.
    kf61: '\E[1;4P', '\EO4P'.
    kf62: '\E[1;4Q', '\EO4Q'.
    kf63: '\E[1;4R', '\EO4R'.
    kind: '\E[1;2B', NULL.
    kri: '\E[1;2A', NULL.
    mc0: '\E[i', NULL.
    mc4: '\E[4i', NULL.
    mc5: '\E[5i', NULL.
    rmacs: '\E(B', '^O'.
    rmcup: '\E[?1049l', '\E[2J\E[?47l\E8'.
    rmm: '\E[?1034l', NULL.
    rs1: '\Ec', NULL.
    rs2: '\E[!p\E[?3;4l\E[4l\E>', '\E7\E[r\E8\E[m\E[?7h\E[?1;3;4;6l\E[4l\E>\E[?1000l\E[?25h'.
    sgr: '%?%p9%t\E(0%e\E(B%;\E[0%?%p6%t;1%;%?%p5%t;2%;%?%p2%t;4%;%?%p1%p3%|%t;7%;%?%p4%t;5%;%?%p7%t;8%;m', '\E[0%?%p6%t;1%;%?%p2%t;4%;%?%p1%p3%|%t;7%;%?%p4%t;5%;m%?%p9%t\016%e\017%;'.
    sgr0: '\E(B\E[m', '\E[0m\017'.
    smacs: '\E(0', '^N'.
    smcup: '\E[?1049h', '\E7\E[?47h'.
    smm: '\E[?1034h', NULL.

Where you see NULL, Konsole did not implement the feature at all. In particular, it did nothing for initc (the feature you are asking about).

Further reading:

• Comparing versions, by counting controls (xterm FAQ)

Answer 2

Seemingly relevant excerpt from man init_color:

The init_color routine changes the definition of a color.  ... The first
argument must be a legal color value; **default colors are not allowed here**.
(See the section Colors for the default color index.)
...
Colors
...
In <curses.h> the following macros are defined.  These are the default colors.
...
             COLOR_BLUE

So these lines are doing something that shouldn't work according to the documentation, but might and do work in some terminals:

int background = COLOR_BLUE;
assert(init_color(background, 1000, 0, 0) == OK);
int foreground = 2;
assert(init_color(foreground, 0, 0, 0) == OK);

There should be 16 predefined color pairs (if we count bold colors), so using pairs 16 and 17 instead of COLOR_BLUE and 2 might work.

(Don't have Konsole to check this, but this might be the reason or at least is worth pointing out.)

Answer 3

Did you mean to set background to COLOR_BLUE? Despite checking can_change_color, it appears that init color doesn't have an affect on Konsole, and the background is staying COLOR_BLUE at init_pair.

int background = COLOR_BLUE;
assert(init_color(background, 1000, 0, 0) == OK);
...
assert(init_pair(pair, foreground, background) == OK);
assert(wattron(window, COLOR_PAIR(pair)) == OK);