Closed MartinNowak closed 9 years ago
The test order is determined by the server, and matches the order of the build which comes from the order of repositories. That code is purely for testing the test-driver.
Additionally, most pull requests either pass completely (executing all tests) or fail at the merge step (which is super quick). Few fail the actual code under test. Rearranging this has little impact on total test throughput.
The test order is determined by the server, and matches the order of the build which comes from the order of repositories. That code is purely for testing the test-driver.
I know, but the driver-v5.sh doesn't seem to match the API code, are they out of sync?
The local script doesn't work, because the arguments are wrong, see https://github.com/MartinNowak/d-tester/commit/5c03d7b6246d62ec63ac4e1d7becbf5a2ed99286.
Also detectos would returns Linux_64
for me, which is an invalid OS according to setup_env.sh.
What versions of the scripts are you actually running?
The client lives in the at-client repo. The server has lots of unpunished changes, but little if any changes in the app, if I remember right. Not in a position to check right now.
On April 17, 2015, at 10:51AM, Martin Nowak notifications@github.com wrote:
The test order is determined by the server, and matches the order of the build which comes from the order of repositories. That code is purely for testing the test-driver.
I know, but the driver-v5.sh doesn't seem to match the API code, are they out of sync?
The local script doesn't work, because the arguments are wrong, see MartinNowak@5c03d7b. Also detectos would returns Linux_64 for me, which is an invalid OS according to setup_env.sh. What versions of the scripts are you actually running?
— Reply to this email directly or view it on GitHub.
I just pushed the server side code changes that have been accumulating for a long time to github, so they should reflect what's running in production.
Thanks, that clarifies how things work ;).