Run integration tests on QNX via QEMU - #328
Conversation
License Check Results🚀 The license check job ran with the Bazel command: bazel run --lockfile_mode=error //:license-checkStatus: Click to expand output |
|
The created documentation from the pull request is available at: docu-html |
This comment was marked as outdated.
This comment was marked as outdated.
| - bazel-config: x86_64-qnx | ||
| bazel-test-target: >- | ||
| //score/... | ||
| bazel-test-target: //examples/... //tests/... |
There was a problem hiding this comment.
It would be good to merge the unit tests and other tests into a single entry so that we utilize the cache more.
There was a problem hiding this comment.
Right now they need to be run with separate configs so that would have to be resolved first.
Our Bazel toolchain for Python does not work with QNX, so we need to use built-in libraries to parse the results instead.
With proper use of Bazel platforms rather than configs, I think it would be possible to do this in a much cleaner way. However that would require deeper changes across our codebase and possibly in the test frameworks as well.
| f"LCM did not exit cleanly, it died with code {proc.get_exit_code()}" | ||
| ) | ||
| assert res == 0, "Couldn't kill lcm" | ||
| proc.stop() |
There was a problem hiding this comment.
Can we still check the exit code here?
In the past I had some test where LCM did crash during shutdown and the test still showed success before the exit code got checked.
| test:x86_64-qnx --test_lang_filters=cc,rust | ||
| build:x86_64-qnx --define config=x86_64-qnx | ||
|
|
||
| build:unit-tests-x86_64-qnx --config=x86_64-qnx |
There was a problem hiding this comment.
Could you extend the Readme at "Getting Started.Building" to show how to run the unit test and integration tests for the different targets?
Might be good to document if we are deviating from the default config.
Closes #316.
Outstanding issues:
--configargument is now needed to run unit tests, as otherwise the unit test framework adds an unwanted second layer of emulation to the integration tests. Since they are using--run_underit is difficult to avoid this.Some of these could be tackled in future PRs.