Why not merge ucs4?

[svenstaro]

The branch seems to work great and reduce SLOC, why not merge it? It's running for me quite well and other people also seem to be using it without issues.

[krypt-n]

I dont really want to diverge further from LemonBoy's master branch, the ucs4 branch introduces a lot of new incompatibilites that will make it harder to merge changes from upstream.
Perhaps the correct way to continue would be to abandon all hope of compatibility in this branch, that would allow us to remove a lot of complexity.

[svenstaro]

Well, did lemonboy show any interest in merging? Or maybe even stated that he won't ever merge this? In that case it would be reasonable to screw compatibility and just do your own thing.

[krypt-n]

I talked with lemonboy about the text encoding. UCS4 brings no advantage for his implementation (the x server is not able to use it), it would add a new dependency (the FcUtf8ToUcs4 function) and implementing codepage conversions on our own would be too much work.

I'll look into removing the server-side rendering in this branch and merging the ucs4 support when I have some spare time