X11 input methods don't appear to work in example programs

[esthermations]

Although it appears Winit has some code to deal with XIM, I can't get my IME to work with any of the example programs.

I should think that XIM would "just work" with winit-based programs, but I don't know. Is it expected that the examples ignore XIM, or is this a bug?

[Determinant]

Got the IME popped out. Should use XSetLocaleModifier before creating an IM instance. The modifier should be the desired IME name such as @im=fcitx. You could manually use other API (such as Xutf8LookupString) from XIM to pull out the strings to received chars.

However, although winit has implemented Xutf8LookupString, it always returns zero when using IME. The reason is now unclear.

[Determinant]

Problem solved. See this issue for how to use an IME in winit.

[Ralith]

@Determinant: I've been investigating why XCompose doesn't work correctly out of the box (I get mojibake instead of the expected text output, e.g. 'Î' and '»' instead of only 'λ'). Your example in #204 was very helpful; it appears that setting libc::setlocale(libc::LC_CTYPE, b"\0".as_ptr() as *const _); is necessary. Can you provide any insight as to why exactly this is necessary? My LANG is set to en_US.UTF-8 and locale reports the same for every variable save LC_ALL (which is empty), and explicitly setting those LC_CTYPE in my environment has no effect, even if I set it to empty. Other X11 programs (e.g. xterm, termite) work fine out of the box, so it seems like winit is probably doing something wrong.

If calling setlocale on startup is necessary for XIM to work at all, I'd like to have winit take care of that automatically--this really should work out of the box--but it doesn't make any sense to me that this is all there is to it. Xutf8LookupString is supposed to return UTF-8 regardless of locale, and my locale is a UTF-8 locale anyway!

edit: setlocale(3) states that the C locale has to be explicitly set to the empty string for the environment variables to be used at all, and setting en_US.UTF-8 fixes things as well, but that doesn't explain why Xutf8LookupString is doing the wrong thing when it's clearly documented as returning UTF-8.