throw if no test items found

[nickrobinson251]

It's always seemed slightly odd that running 0 tests is a "success", in #215 we at least made it look less successful by logging a warning, but ultimately this doesn't catch issues (like we've seen at RAI) where e.g. changes in CI test setup leads to some job running 0 tests and therefore the job succeeding and therefore the issue (that we're no longer running tests) being missed.

If you runtests and fail to run any tests, i think that should be an error. So this PR makes it an error.

Here's how it looks:

julia> runtests("test/testfiles/_happy_tests.jl"; name="blahahah_nope")
[ Info: Scanning for test items in project `ReTestItems` at paths: test/testfiles/_happy_tests.jl
[ Info: Finished scanning for test items in 0.04 seconds.
ERROR: No test items found.

Note we follow the precendent of Test.TestSetException and print just the ERROR: ... line without a subsequent backtrace

e.g. compare this where there's no backtrace after the final Error: ... line

julia> @testset "foo" begin
           error("barf")
       end
foo: Error During Test at REPL[31]:1
  Got exception outside of a @test
  barf
  Stacktrace:
    [1] error(s::String)
      @ Base ./error.jl:35
    ...snip...
   [16] _start()
      @ Base ./client.jl:552
Test Summary: | Error  Total  Time
foo           |     1      1  0.0s
ERROR: Some tests did not pass: 0 passed, 0 failed, 1 errored, 0 broken.

julia>

This PR also updates the error when the file doesn't exist (when validate_paths=true, or no valid files were given, resulting in no testitems):

julia> runtests("no_such_tests.jl"; validate_paths=true)
ERROR: No such path "no_such_tests.jl"

[Drvi]

Makes sense, thanks!