This commit adds a proof-of-concept tool for combining log files from multiple instances of bitcoinds as well as the test_framework.log file when running a functional test. This gives a combined view of what the test framework and all bitcoin instances were doing during the test.
There are 3 options for output: plain text, colorized text for posix terminals, and html:
Plaintext:

Colorized text:

HTML:

Possible future enhancements:
- add some very simple javascript to the HTML output to allow filtering of log events by source, severity, timestamp, regex, etc.
- have the test_runner generate combined logs automatically if a test fails.
- have automated build tools like travis to generate and push an html log of a testcase if it fails
