Dockerfile: Install sparse package for code analysis

[aiChaoSONG]

The sparse tool can be used to do static code analysis,
which helps to find coding faults at compile-time.

Signed-off-by: Chao Song chao.song@linux.intel.com

The Sound Firmware Team will need this package for sparse test. I think zephyr should do sparse test now or sooner.

We have to clone the latest code and compile, the sparse package shipped with ubuntu 20.04 can not work for us.

@lyakh @keqiaozhang FYI

[aiChaoSONG]

I see this error, Error: The log was not found. It may have been deleted based on retention settings. can someone tell me what happened, how can i fix this?

[cjgraydon]

Looks like the CI workflow for the PR failed due to some kind of runner/network issue: https://github.com/zephyrproject-rtos/docker-image/actions/runs/2588361237

i.e. Build
The hosted runner: GitHub Actions 15 lost communication with the server. Anything in your workflow that terminates the runner process, starves it for CPU/Memory, or blocks its network access can cause this error.

It looks like the run timed out and it never made it far enough to generate the log that subsequent workflow steps want. The step before where you see the log error message threw some kind of error.

You might try updating your PR or doing an empty commit to try to trigger another run: git commit --allow-empty -m "this is an empty commit" Pushing that should force CI to run again, in theory. Might just be something flukey that caused the initial run to time out.

Looks like stephanoslo might have attempted a couple of re-runs, though. So might not be so simple. Just something to try, for what it is worth.

[aiChaoSONG]

@cjgraydon Thank you, but looks I am first-time contributor, need approval to run GH action.

[aiChaoSONG]

failed at Installing uefi-run utility, not even touch my changes.

[stephanosio]

It looks like the resources available on the default GitHub-provided CI runner is no longer sufficient to build the docker image. I will see if I can rework the CI to fix this.

[cjgraydon]

@cjgraydon Thank you, but looks I am first-time contributor, need approval to run GH action.

Side note: Any additional push to your PR should trigger a new CI run. So you could dodge having to have direct access to the GitHub actions. I'd think an empty commit to your PR would do the trick, but I might be off on that. I use this approach in our enterprise GH on occasion. Might be different in public GH, though? Anyway, just a little more detail on what I was talking about in case I wasn't especially clear. :)

[marc-hb]

A record of which particular sparse SHA1 was used must be kept somewhere. I don't know how that's normally done, maybe you can add a new ARG SPARSE_COMMIT variable at the top and check that out?

[aiChaoSONG]

@stephanosio Is CI good now? please help to approve and test this patch, thank you.

[stephanosio]

@stephanosio Is CI good now? please help to approve and test this patch, thank you.

sorry, I haven't had time to fix the Docker image CI yet (apparently the problem turned out to be not so simple to fix, will likely need an overhaul ...).

I will let you know when it is fixed and merge this PR thereafter.

[aiChaoSONG]

@stephanosio Thank you, the test passed, will you push a new image to docker hub after this merged?

[stephanosio]

unless you actually need a feature from the unstable/development branch, we should use a release tag.