Drop support for LCMS version 1

[aclark4life]

Due to security issue: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=718682

[wiredfool]

Ok. Take 1. https://github.com/wiredfool/Pillow/tree/lcms2

Compiles, runs, some tests fail, I think only the ones that deal with attributes.

There's some impedance mismatching from the icc v2 ->v4 profiles (or maybe it's just lcms1->2) in the text attribute handling portion. Tests are failing there, likely because I'm reading the wrong header field. That's one section that's been greatly reorganized. I'd like to support the unicode version of the attribute call, but I need to work around the wchar_t buffer that's needed there.

[wiredfool]

Ok. Done a bit of digging on the info and name fields, and lcms1 combined a bunch of information into those, and didn't leave an efficient way to get them in lcms2 (at least, that I've found). The name isn't too bad to backfill, the info is tricky, since it calculates a white balance and does some other stuff with a private kodak tag.

I've exposed the 4 main tags, model, manufacturer, copyright and description. I've fixed up the name. And I've got no idea where some of the other data for the tests is coming from (like the (lcms internal) item. I suspect that might even just be in lcms1, as it's a built in profile.

Anyway, there are 3 tests failing, and I think they're minor, and possibly worth just fixing the test instead of the code. I'd love to hear from users on that.