Add option to set output file

[nprizal]

Description

in #156, we changed how we trigger the retry by saving the output of Rspec run to detect failed tests. The output of Rspec run was saved in a temporary file that will be deleted at the end of operation.

This PR adds a new option called BUILDKITE_SPLITTER_RESULT_PATH that can be used to specify the path to save the output of Rspec run. If the configured path has *, it will be replaced with a UNIX timestamp, this is useful when the retry function is enabled, and we want to store the output for each run separately without overriding the previous run.

Testing

I've done integration testing manually

• set BUILDKITE_SPLITTER_RESULT_PATH=./tmp/my-rspec-result-*.json
• run ./test-splitter
• check if a json file exists in $BUILDKITE_SPLITTER_RESULT_PATH configured above

[nprizal]

What happens if the configured path doesn't have * and retries are enabled? Does that mean it won't work properly?

@niceking It still works properly, because we interpret the report before we execute the retry runs. It just means that the result from retries will override the result from previous run(s).