Hello!
I've been trying to implement Unicode keyboard input in my regular lite-C Pure game.
Using on_message I receive WM_CHAR events, including characters sent by a Japanese IME thing for testing. However, it seems like I cannot convert those to actual Unicode characters (the lParam is just 1 which probably is not right for any subsequent ToUnicode calls...?).
I assume this might be because IsWindowUnicode(hWnd) returns 0 and WM_CHAR messages are transformed to ANSI before they are sent to the application.
I've thought about creating my own hWnd with RegisterClassExW (which is not supported in GStudio8/include/windows.h, only RegisterClassEx, so maybe via a DLL) but I couldn't figure that one out yet. Would that be the right way?

Alternatively, I've tried to call
SetWindowsHookEx(WH_KEYBOARD_LL, HookCallback, NULL, 0);
in a DLL or directly in lite-C, then do the Unicode translation there and send the Unicode characters to the engine.
However, it seems like the callback function is not triggered correctly, in particular no response to/ from IME input. WH_KEYBOARD_LL may or may not be the wrong HookID but I did not have luck with any other mode.

Do you have any advice?
Thanks for reading.

I gave it a shot and jap. IME does indeed not trigger KEYDOWN events. You might need some other functions to retrieve IME text since it's not really a key press/release but a string of characters that's sent to the window.

Yes, the hook is destroying the inputs. (EDIT: Oh, and thanks for giving it a shot!)

I've created a new Unicode hWnd window with the same/ default on_message engine message loop. That one is receiving the same messages as the default engine window, so the issue must be somewhere else (in the message loop itself?).

I don't think I can do this on my own.
Mr. Lotter, do you have any advice?

Thanks.

Anyone got an idea how to get Unicode input running in A8?

Mr. Lotter, can you do something about this, if necessary, release an official/ inoffical interim version with Unicode support? That would be amazing!

The issue in short: on_messages receives only ANSI WM_CHAR input (for IME input which is received as a sequence of multiple WM_CHAR messages, which is correct, this means they oftentimes all have the same values although it's different characters).

So you never actually catch a WM_UNICHAR message, only WM_CHAR?

https://msdn.microsoft.com/de-de/library/windows/desktop/ms646288(v=vs.85).aspx

See comment concerning wParam: Sounds like what is happening to you.

Then you could build your own inkeyw based on custom message handling.

Thanks for the suggestion, I actually hadn't found WM_UNICHAR yet, always only looked at WM_CHAR or non-filtered input. :-X
However, I cannot seem to trigger WM_UNICHAR at all.
I found WM_IME_CHAR which gets fired N times before N characters from an IME arrive, however they are converted to ANSI as far as I can tell before I get my dirty fingers on them. Even when I CreateWindowExW() a new Unicode window (IsWindowUnicode() returns 1 then) I cannot extract or combine WM_CHAR events into unicode digits.

So if you add your own lessage handling and WM_IME_CHAR message fires, what does wParam contain then? Is this value already truncated?

wParam contains a bunch of values, most often 63. Currently I'm only receiving a single WM_CHAR event for each Unicode character - I thought I could trigger multiple calls some time ago -, all of them in the 8 bit range.

Hm... 63 is a '?' .... supsicious somehow.
All values for wparam are always in 8 bit range? Strange.

According to the msdn you may receive both single and double byte characters, but this is not Unicode then and requires conversion - which is not trivial at all.