Test unicode use and encoding awareness

[svanoort]

We need better test coverage of Unicode functionality. This ties into Python3 compatibility, because otherwise that enhancement has high potential to break things. This would be a very easy item for a new contributor to add to!

Places to add:

• test_tests.py -- test of test parsing here
• contenthandler.py -- test_contenthandling.py (use a tempfile for testing handling of file read)
• functionaltest.py -- actual running test with a local DB (CRUD operations)
• content-test.yaml -- do a basic CRUD test of content Created unicode-test.yaml, adding in section at a time

Components:

• Fix to url handling to better concatenate/encode unicode characters
• Test that request bodies work right with different encodings when: (test_contenthandling.py)
  • Included inline in YAML
  • Included inline in YAML with template applied
  • Read from file - N/A
  • Read from file with templating applied - N/A, requires In contenthandler, allow specifying an encoding for file content #121 for Unicode file templating
• Test that URLs correctly do unicode & special character URL encoding, with and without templating -- this turns out to be amazingly complete/painful, left some tests in place and opened Support internationalized URIs (unicode URLs) - RFC3986 #123 for the URI/IRI mapping bits
• Test that validators/extractors/etc can work with unicode content - done, json parser is thankfully unicode-smart, whee.

* Internally insure that string data is stored as unicode and encoded into raw bytes at the LAST moment
* functional tests for unicode use - Tests are actually passing correctly

[svanoort]

@alexeyknyshev This would be a great place to contribute, since I know you're already looking at how unicode works with PyRestTest!

[svanoort]

Unicode handling policy:

• Internally all string values are converted to Unicode for consistency at the first opportunity
  • For YAML reading, they're unicode anyway (UTF-8 is the YAML encoding AFAICT)
• Binary content is kept as bytes where necessary (example: request/response bodies)
• PyCurl will accept unicode containing only ASCII character points, any escapes/encoding to bytes are done at the last possible second when configuring PyCurl itself
• All operations should be Unicode-safe (templating included, via helper methods to do un-encode and re-encode operations since native string.Template doesn't allow Unicode)

EDIT:
URL handling is tricky, pipeline needs to go like this:
URL base + url --> templating (yay) --> URL encoding if contains non-ASCII characters?

[svanoort]

Completed as of 2899269 (plus previous commits in branch)