R1:6eea500bbd8b

Document character encoding policies and how to fix mangled UTF8 files

Summary:
See D431, where I promised to document this like 2 months ago. Document that:

- Everything should be UTF-8.
- ASCII is recommended.
- How to identify and repair files which aren't valid UTF-8.
- What to do if you're using some other encoding.

Test Plan: Generated and read documentation.
Reviewed By: codeblock
Reviewers: edward, codeblock, jungejason, tuomaspelkonen, aran
CC: aran, codeblock
Differential Revision: 776