[Solved] Ncurses and unicode

developer94 · 2014-06-25 14:37:14

Hey everyone,

I'm having a bit of an issue with understanding Allan's idea of having a single ncurses header/library(?) for both ASCII compatible and widechar text.
Since ncurses.h doesn't provide functions like add_wch/addwstr or other wide character manipulation functions, I can't seem to figure out a way to do it.

This is what I've got so far. All I want to do is find a way how to output a darn unicode checkmark.

#define _X_OPEN_SOURCE_EXTENDED
#include <stdio.h>
#include <locale.h>
#include <wchar.h>
#include <ncurses.h>

int main(int argc, char **argv)
{
    setlocale(LC_ALL, "");

    initscr();
    cbreak();

    refresh();

    mvaddch(0, 0, 'a'); // let's make sure it does output
    mvaddstr(1, 0, L"\x2713"); // then let's output the checkmark

    refresh();

    while(1);

    endwin();
    return 0;
}

Can anyone provide info on how single-header ncurses handles unicode, or a short example perhaps?

Cheers,
Johnny

Last edited by developer94 (2014-06-26 08:31:20)

Cloudef · 2014-06-25 14:53:04

Don't use L"", don't use wide chars, and link to ncursesw. (addwstr) etc are not wide char version of the functions, but they take WINDOW as argument.

developer94 · 2014-06-25 15:50:34

Thanks for the ideas!

But, if I do not use L"", I'm in for an even bigger mess. 0x2714 doesn't fit in a byte, and the compiler warns me about that aswell.
I am linking to ncursesw, so that shouldn't be an issue.
Also addwstr should be a widechar variant of addstr. I think you might be refering to waddstr which takes WINDOW* as argument.

Any other clues?

rockin turtle · 2014-06-26 05:39:23

I would try encoding the check mark in utf-8, which doesn't use wide characters.

The checkmark at unicode codepoint 0x2713 would encode to the three bytes e2 9c 93 in utf-8.

Codepoint 0x2713 -> e2 9c 93
Codepoint 0x2714 -> e2 9c 94

	mvaddstr(1, 0, "\xe2\x9c\x93");

see http://en.wikipedia.org/wiki/UTF-8 for more info on utf-8.

Cloudef · 2014-06-26 07:00:56

developer94 wrote:

Also addwstr should be a widechar variant of addstr. I think you might be refering to waddstr which takes WINDOW* as argument.

Yeah, you are right about addwstr, the naming conventions in ncurses seems to differ from function to function.

I am linking to ncursesw, so that shouldn't be an issue.

You should use UTF8 like rockin turtle advised.

developer94 · 2014-06-26 08:30:57

Jeeee, what was I thinking!
I went too desparate in finding a solution that I overlooked set_locale setting it to the default locale (utf-8 in my case).

Should the wiki be expanded to note this pecularity (the lack of direct widechar support in ncurses)? I could do it, but I'm new and ain't really sure how does stuff work around here.

Thanks for the help fellas!

Arch Linux

#1 2014-06-25 14:37:14

[Solved] Ncurses and unicode

#2 2014-06-25 14:53:04

Re: [Solved] Ncurses and unicode

#3 2014-06-25 15:50:34

Re: [Solved] Ncurses and unicode

#4 2014-06-26 05:39:23

Re: [Solved] Ncurses and unicode

#5 2014-06-26 07:00:56

Re: [Solved] Ncurses and unicode

#6 2014-06-26 08:30:57

Re: [Solved] Ncurses and unicode

Board footer