#308 put CaM sources into zip archive

[Raleksan]

Work under #308: Adding CaM sources to a zip archive.

Some notes:

• The path to clone the repo is ${TARGET}/github/self/${name} to use a common command for all repos, removing the .git directories.
• A user may be using an out-of-date version of CaM, in which case they will keep an archive with a CaM repo of a newer version.

[yegor256]

@Raleksan what about a unit test for this new feature? Simply put, how do you know that it works?

[Raleksan]

I think adding a test for this feature is achievable for me.

As I understand, the tests for the scripts in the /steps directory are located in /test/steps and to add a new test I need to edit test/steps/test-zip.sh.

Am I getting this right?

p.s. Do I always need to ping you with @*? Or will you get the notification anyway?

[yegor256]

@Raleksan you are right, the tests are in the test/steps/test-zip.sh. Also, read this please: https://www.yegor256.com/2024/04/01/ping-me-please.html

[Raleksan]

@yegor256 I added a new test, you can review it.

[yegor256]

@Raleksan I believe, it is reasonable to delete the "${cam_repo_target_dir}/.git" directory after the clone

[Raleksan]

@yegor256 I know that I need to delete .git, therefore I locate it at ${TARGET}/github/self/${name} to delete .git directory with all others git directories, using already existing command, or it is better to do it separately and explicitly?

[yegor256]

@Raleksan this location is wrong, I believe. Let's put it into ${TARGET}/cam-sources instead

[Raleksan]

@yegor256 I changed the location as you remarked and improved the test.

Can you please review it again?

[yegor256]

@Raleksan not all tests pass after your changes. Try to run make test locally.

[Raleksan]

@yegor256

I have analyzed the pipeline run errors and found that both failures depend on the same test test-integration.sh.

I checked out test-integration.sh and here is my interpretation, could you please clarify?

The actions of test-integration.sh:

• Initially it uses https://github.com/yegor256/tojos as a single repo to collect metrics (lines 29-44).
• After analysis, it checks that all artifacts (hashes, reports, etc) exist (lines 45-48)
• Checks the number of entities contained in the TARGET directory [this step raises an error] (line 49)
• Checks if certain files exist (lines 50-66)

I'm interested in line 49:

test ‘$(find “${TARGET}” -maxdepth 1 | wc -l | xargs)’ = 10.

It's unclear to me from where we inherit this ‘magic number’.

For this PR, I added 1 entity to TARGET which is CaM sources, which I think is causing the error in this test.

p.s. These are just my thoughts without running tests because I faced a bug while building a Docker container locally, I'm going to investigate and try to fix this problem or open an issue if if it's a real bug

[yegor256]

@Raleksan here, 10 is the number of highest-level elements we expect to have in the array. It's indeed a magic number. I believe, you should just change it to 11 and the test will pass.

[Raleksan]

@yegor256 I corrected this value, can you run CI to test it? I'm still having problems with the local build of the Docker image.

[yegor256]

@Raleksan if you have problem with Docker, please report an issue, we will try to resolve it

[yegor256]

@Raleksan thanks!