Some executors do not run the tests in a shell, and so assuming that
they can understand shell builtins is wrong. Instead, call Bash
directly to do what we need to do.
This still requires the executor to be able to run Bash, but at least
it does not require it to interpret command lines in the Bash language.
Instead of passing file dependencies individually, assume that the
whole content of the unique test directory is a dependency. This
simplifies the test harness significantly, by making %T the directory
that contains everything required to run a test. This also removes the
need for the %{file_dependencies} substitution, which is removed by this
patch.
Furthermore, this patch also changes the harness to execute tests locally
inside %T, so as to avoid creating a separate directory for no purpose.