I think it would be cool to provide a test runner per implementation dir e.g. in js/jsedn/runner I will have:
- install.sh file that just runs npm install (but could potentially do other things if necessary)
- run-valid-tests.sh which runs all of the valid-edn tests.
- run-invalid-tests.sh which runs all of the invalid-edn tests.
- run-performance.sh which runs the performance tests.
All run-*-tests scripts will provide standard output (an edn map actually keyed by filename - for valid and invalid this will be true or false for values and for performance it will be microsecs test took as values) so it can be tested against other implementations.
I think this will be nice and platform agnostic whilst allowing each implementation to do what it needs setup wise.
Also this should make it easy to have an overall runner that runs all of them and collates the output (which I want to use to generate a nice gh-pages with).
I think it would be cool to provide a test runner per implementation dir e.g. in js/jsedn/runner I will have:
All run-*-tests scripts will provide standard output (an edn map actually keyed by filename - for valid and invalid this will be true or false for values and for performance it will be microsecs test took as values) so it can be tested against other implementations.
I think this will be nice and platform agnostic whilst allowing each implementation to do what it needs setup wise.
Also this should make it easy to have an overall runner that runs all of them and collates the output (which I want to use to generate a nice gh-pages with).