Improve run-test tooling support
A useful thing to do with run-test tests is to run them in parallel to
reproduce low-probability flakes. We have a tool called
parallel_run.py to do this. Unfortunately one had to manually create
the script that parallel_run.py would execute, making it rather
difficult to use. This adds a --create-runner flag to run-test that
will cause it to write a script (called $DEX_LOCATION/runit.sh) that
is compatible with parallel_run.py
Test: % ./test/run-test --host --64 --prebuild 001-HelloWorld
...
Runnable test script written to /tmp/allight/test-180018/runit.sh
...
% ./tools/parallel_run.py /tmp/allight/test-180018/runit.sh -j80
Change-Id: I18d4ec72b51fb3c868ca4c6fe7f191e74f45b868
2 files changed