Improve Auto-Naming Test methods

[markcmiller86]

• Have them take section headings and/or file prefixes from the funciton's doc string instead of the function's name.
• Instead of using an incrimenting index for tests from within the same function, find a way to create a hash based on a more permanent expression of the test (key VisIt CLI methods). For example, from this block of code, https://github.com/visit-dav/visit/blob/8f70eeda2bd57748a22f119d55a70bbb229bfe29/src/test/tests/queries/line_scan.py#L82-L102

    OpenDatabase("cld_i0.ult") # test-tag
    ReOpenDatabase("cld_i0.ult")
    AddPlot("Curve", "Chord length distribution - individual") # test-tag
    DrawPlots()
    ResetView()
    test1 = "ls_%d_%d" %(index,1)
    TestAutoName(test1)
    Query("Integrate") # test-tag
    text = GetQueryOutputString() # test-tag
    test2 = "ls_%d_%d" %(index,2)
    TestTextAutoName(test2, text)

The auto-naming methods could gobble up and hash all preceding lines with # test-tag in them (whitespace, etc. removed) and use that as the test's index. In the example above. This would make the auto-naming logic less sensitive to minor structural changes in the python code. For example, a new test could be added between the existing two without causing them to rename.