Bash test results are stored in test/*/test-64-results. They are then filtered
by COMPONENT_TEST_TRANSFORMS to make the test produce stable/reproducible
results. Currently the only test which is not sometimes passing is 'run-trap'.
Explanation is in the test output itself, and this is what the output looks like:

  warning: UNIX versions number signals and schedule processes differently.
  warning: If output differing only in line numbers is produced, please
  warning: do not consider this a test failure.
  71d70
  < +[8] cat
  72a72
  > +[8] cat


COMPONENT_TEST_TRANSFORMS thus remove this test output. When testing bash it is
necessary to observe full test-64-results.


Please note that the tests have to be executed from tty, otherwise this two
tests fail:

run-execscript
run-read


To test new version of bash installed on new machine I did:

- install new bash
- reboot
- svcs -x # to see if it caused any issue
- bash --version
  - to see if we are testing right bash version