Converting scancodes to ASCII - c++

Question

I'm implementing my own text editor in c++. It's going... ok. ;P
I need a way to turn a keycode (specifically Allegro, they call it scancodes) into an ASCII-char. I can do A-Z easy, and converting those to a-z is easy as well. What I do currently is use a function in Allegro that returns a name from a scancode (al_keycode_to_name), meaning if the key pressed is A-Z it returns "A" to "Z". That's easy peasy, but I can't simply read special characters like ",", ";" etc. That's where I'm having a hard time.
Is there a way to do this automatically? Maybe a library that does this? The real trick is taking different layouts into consideration.
Here's what I have so far, in case anyone's interested. The class InputState is basically a copy of the Allegro inputstate, with added functionality (keyDown, keyUp, keyPress for example):
void AllegroInput::TextInput(const InputState &inputState, int &currentCharacter, int &currentRow, std::string &textString)
{
static int keyTimer = 0;
static const int KEY_TIMER_LIMIT = 15;
for (int i = 0; i < 255; i++)
{
if (inputState.key[i].keyDown)
{
keyTimer++;
}
if (inputState.key[i].keyPress)
{
keyTimer = 0;
}
if ((inputState.key[i].keyPress) || ((inputState.key[i].keyDown) && (keyTimer >= KEY_TIMER_LIMIT)))
{
std::string ASCII = al_keycode_to_name(i);
if ((ASCII.c_str()[0] >= 32) && (ASCII.c_str()[0] <= 126) && (ASCII.length() == 1))
{
textString = textString.substr(0, currentCharacter) + ASCII + textString.substr(currentCharacter, textString.length());
currentCharacter++;
}
else
{
switch(i)
{
case ALLEGRO_KEY_DELETE:
if (currentCharacter >= 0)
{
textString.erase(currentCharacter, 1);
}
break;
case ALLEGRO_KEY_BACKSPACE:
if (currentCharacter > 0)
{
currentCharacter--;
textString.erase(currentCharacter, 1);
}
break;
case ALLEGRO_KEY_RIGHT:
if (currentCharacter < textString.length())
{
currentCharacter++;
}
break;
case ALLEGRO_KEY_LEFT:
if (currentCharacter > 0)
{
currentCharacter--;
}
break;
case ALLEGRO_KEY_SPACE:
if (currentCharacter > 0)
{
textString = textString.substr(0, currentCharacter) + " " + textString.substr(currentCharacter, textString.length());
currentCharacter++;
}
break;
}
}
}
}
}

Answer 1

You should be using the ALLEGRO_EVENT_KEY_CHAR event with the event.keyboard.unichar value to read text input. ALLEGRO_EVENT_KEY_DOWN and ALLEGRO_EVENT_KEY_UP correspond to physical keys being pressed. There is not a 1:1 correspondence between them and printable characters.
Say a dead key is being used to convert the two keys e' to é. You'd get two key down events for e and ' (and neither are useful for capturing the proper input), but one key char event with é. Or inversely, maybe somebody mapped F4 to a macro that unleashes an entire paragraph of text. In that case, you'd have multiple chars for a single key down.
Or a simple test: if you hold down a key for five seconds, you will get one ALLEGRO_EVENT_KEY_DOWN but multiple ALLEGRO_EVENT_KEY_CHAR as the OS' keyboard driver sends repeat events.
You can use ALLEGRO_USTR to easily store these unicode strings.
ALLEGRO_USTR *input = al_ustr_new("");
// in the event loop
al_ustr_append_chr(input, event.keyboard.unichar);
There's also ways to delete characters if backspace is pressed, etc. You can use the ustr data types with the font add-on directly via al_draw_ustr(font, color, x, y, flags, input), or you can use al_cstr(input) to get a read-only pointer to a UTF-8 string.