so that tfkit can now be tracked in separate repo.  (In future we might
add it as submodule)

The tool is mostly based on ini files; make this more obvious.

Support test folder structure where test, oracle and generated result
data are stored each in separate directory.  This makes it easier to
ensure legible diff output without cluttering the filenames too much.

Most changes are in tf_do_subtest.

 *   Separated setup/cleanup logic from test data collecting and
     execution/evaluation.

 *   Revamped the execution/evaluation logic in a new tool testflt,
     allowing tester to use it as the need be.

 *   Code to compose actual test command removed, as testflt makes it
     easy to write the code in one place (TF_RUN) and keep simple things
     simple.

 *   Removed tf_parse_meta, as it's obsolered by above changes.

 *   Updated tests to reflect the new style.

Avoiding using the plain test name makes it easier to understand the
role of each file.

This new hierarchy makes it easier to work simultaneously on data and
the script, simplifies deployment to tempdir and makes the launching
script naming consistent with simple.sh.

config.cfgrep does better job at hinting how the function works
(i.e. somehow similar to UNIX grep) and that the use is not limited to
INI files.

Also default config file suffix is now "conf".

As a followup to last revamp:

 *  Use more extensible test definition format,
 *  separate testing framework from setup script and split to includes,
 *  update test scripts as well as data,
 *  consolidate test output,
 *  and make some small improvements/fixes on the way.

 *  Enable running all tests via make,
 *  simplify tests (lay off data/tmpdir manipulation to framework),
 *  set rules for naming, exit statuses, etc,
 *  remove beakerlib dependency.