Make sure FancyReporter detects encoding of the traceback messages.

[sublee]

Hi Dag,

I'm going to use Attest==0.5.3 instead of Unittest for my project focused on Korean language. So my test suites contain many Hangul characters (non-ASCII) and are hinted encoding by # -*- coding: utf-8 -*-. But traceback message of FancyReporter is broken because it doesn't aware the encoding of the test code.

# -*- coding: utf-8 -*-
from attest import Tests, assert_hook

suite = Tests()

@suite.test
def unicode_failing():
    assert u'유니코드' == u'테스트'

The output will be broken:

[100%] 1 of 1 Time: 0:00:00

koreantests.unicode_failing
--------------------------------------------------------------------------------
Traceback (most recent call last):
  File "koreantests.py", line 7, in unicode_failing
    assert u'ì ëì½ë' == u'í
ì¤í¸'
TestFailure

assert (u'\uc720\ub2c8\ucf54\ub4dc' == u'\ud14c\uc2a4\ud2b8')
assert (u'\uc720\ub2c8\ucf54\ub4dc' == u'\ud14c\uc2a4\ud2b8')

Failures: 1/1 (1 assertions)

My patch provides the code encoding detector for fix broken messages of FancyReporter:

[100%] 1 of 1 Time: 0:00:00

koreantests.unicode_failing
--------------------------------------------------------------------------------
Traceback (most recent call last):
  File "koreantests.py", line 7, in unicode_failing
    assert u'유니코드' == u'테스트'
TestFailure

assert (u'\uc720\ub2c8\ucf54\ub4dc' == u'\ud14c\uc2a4\ud2b8')
assert (u'\uc720\ub2c8\ucf54\ub4dc' == u'\ud14c\uc2a4\ud2b8')

Failures: 1/1 (1 assertions)

Please review and accept this pull request if there's no problem.

Regards, Heungsub