Sometimes, the text after "# " shows the name of the test binary
("<package>.test") in addition to the package
e.g.: "# package/name [package/name.test]".
This change ensures that the packages' build outputs are associated
correctly in this case.
Mainly:
* Moving the averaging/merging of benchmarks from the parser to the formatter package
* Tightening up the regex so it only captures the numeric values (no more of trimming spaces and the ns/op)
* Deleting the writing up in xml file the benchmark memory sections of B/op and Allocs/op
Also added a test case for parseNanoseconds().
Whenever we encounter a result line in the test output, we would find
the first test with the matching name and update it. However, in some
cases it's possible for the same test name to appear multiple times in
the output. To prevent us from always updating the oldest test, the
order in which findTests searches for a matching test is reversed so
that it always returns the most recently added test.
Fixes#54.
If a package compiles correctly, but panics before it has a chance to
run any tests it would previously be ignored. Any failing packages
without tests but with some output will now be included in the report
with a dummy test.
Fixes#52
In case of build error, the unittest fail, but the junit-report
with the set-exit-code flag on returns status code 0.
This commit will make it return the right status code.
