Automated tests for flit update

[mikebentley15]

Bug Report

Describe the problem
The flit update command-line tool (i.e. the flit_update.py script) does not have sufficient automated tests. There is a pull request for a new feature in there with automated tests for that feature, but there are existing features that need tests. The list I see now are:

1. FLIT_INC_DIR, FLIT_LIB_DIR, FLIT_DATA_DIR, and FLIT_SCRIPT_DIR, for
   both from git repository and for an installed FLiT
2. DEV_CC, DEV_OPTL, and DEV_SWITCHES for both provided and non-provided
3. GT_CC, GT_OPTL, and GT_SWITCHES for both provided and non-provided
4. TEST_RUN_ARGS from timing, timing_loops, and timing_repeats for both
   provided and non-provided
5. ENABLE_MPI and MPIRUN_ARGS for both provided and non-provided
6. HOSTNAME
7. UNAME_S

Suggested Fix
Write either a large automated test for this or many small automated tests.

Alternative approaches:
Just don't test it as it is now, which isn't a good approach if we want to identify regressions.