Improved unicode handling

[joerick]

Some unicode characters are supported (if they're in the font), but some cause the renderer to crash, for example:

screen.text(u'🕈☟⚐ 💧✌✋👎 ✡⚐🕆 🕈☜☼☜ ✋☠ 👍⚐☠❄☼⚐☹✍ ☟☜☟☜☟☜')

Traceback (most recent call last):
  File "main.py", line 33, in <module>
    screen.text(u'🕈☟⚐ 💧✌✋👎 ✡⚐🕆 🕈☜☼☜ ✋☠ 👍⚐☠❄☼⚐☹✍ ☟☜☟☜☟☜')
  File "/Applications/Tide.app/Contents/Resources/tingbot/graphics.py", line 208, in text
    super(Screen, self).text(*args, **kwargs)
  File "/Applications/Tide.app/Contents/Resources/tingbot/graphics.py", line 150, in text
    text_image = Image(surface=font.render(string, antialias, _color(color)))
UnicodeError: A Unicode character above '\uFFFF' was found; not supported

[joerick]

Looks like a pygame issue, may be fixable by moving to the freetype backend

[furbrain]

I'm trying to reproduce this in linux, but it doesn't seem to cause any problems for me. However, I've only cut and pasted from the above link - not sure how well the unicode has come across.
I've also tried screen.text(u'\U00010143') - which should print a gamma sign but gives me a capital N with a tilde - but no crash.

[joerick]

Interesting. Probably pygame.font uses a different backend on the different platforms. I'd really like emojis to work as well, maybe pygame.freetype would provide that? We'd also have to bundle an emoji font and a unicode font with a good number of glyphs for fallback.