Unicode support? about pancurses

[saethlin]

I added features = ["wide"] to my Cargo.toml, and unicode output works now!
Is there still work to be done? If this feature is incomplete or something I'm interested in helping out.

[ihalila]

Could you elaborate on what you mean exactly? At least on Windows unicode output works as far as I know, aside from some special cases, and input should as well for both platforms with the release of v0.15. Output with ncurses may require more work, I haven't looked into it for a while.

[saethlin]

I'm looking at a problem with unicode output on linux.

I think all one would need is to insert this call after the initscr() call, but I'm a bit hesitant to suggest just making everyone en-US locale.
https://github.com/jeaye/ncurses-rs/blob/85ab2aa9d58eb58ba68cbac47f105cc93cf36aae/examples/ex_7.rs#L19

[ihalila]

setlocale is already called in

pancurses/src/unix/mod.rs

Line 15 in 65621d5

setlocale(LcCategory::all, "");

, so that's not the problem.

I did some testing and Unicode output works as it should for me as long as I have the ncurses/wide feature enabled (exported as just wide in pancurses) and I don't use addch, because that doesn't work correctly. Checking the ncurses manual at http://invisible-island.net/ncurses/man/curs_addch.3x.html it has two relevant bullet points:

• check if a character can be represented as a single byte in the current locale before attempting call waddch, and
• call wadd_wch for characters which cannot be handled by waddch.

which leads me to believe that pancurses should really always use wadd_wch, it would probably work correctly. Unfortunately ncurses-rs does not expose that function yet so fixing this requires a change there first.

[ihalila]

The only issue I know of is the one I described in my earlier comment, regarding addch() not working for multibyte characters.

[TimonPost]

There is one other problem as described here:

#43 (comment)

We can't use Unicode on a terminal created with pancurses::newterm. In order to support this platform_specific::pre_init() should be added to that function as well.

[plefebvre91]

Is there any solution today ?