build-and-test-differential-above fails do to missing dependency to autoware_system_designer

[mitsudome-r]

Checklist

• I've read the contribution guidelines.
• I've searched other issues and no duplicate issues were found.
• I'm convinced that this is not my fault but a bug.

Description

build-and-test-differential-above fails do to missing dependency to autoware_system_designer.
Examples:

• https://github.com/autowarefoundation/autoware_universe/actions/runs/25223454355/job/73961133809?pr=12534
• https://github.com/autowarefoundation/autoware_universe/actions/runs/24873773551/job/72825820765?pr=12496

Expected behavior

Workflow passes without failure

Actual behavior

Workflow is failing

Steps to reproduce

Create a package that modifies packages that trigger the build of autoware_system_designs package in autoware_launch (e.g., autoware_pointcloud_preprocessor).

Versions

No response

Possible causes

Why it fails

packages_above.repos clones launcher/autoware_launch into the workspace, which contains autoware_sample_designs. Therefore, whenever there is a package that is depended by autoware_sample_designs is modified (like autoware_pointcloud_preprocessor), the workflow fails with the error.

How to fix

I can think of the following options to fix the error:

1. Add autoware_system_designer to build_depends_above.repos or build_depends_stable. Ensures the workspace builds it from source matching a pinned tag.
2. source /opt/autoware/ since autoware_system_designer should be already built within ghcr.io/autowarefoundation/autoware:universe-dependencies-* images used for the ci.

I prefer to go with option 2 so that we don't have to rebuild the package which are already built. One drawback is that when ever we want to update the package version, we would have to wait for the docker build pipeline in autowarefoundation/autoware to finish, but I don't think we would expect frequent change for the repo.

Additional context

No response

[mitsudome-r]

After discussion in the draft PR, I decided to fix the error with approach 1 for now.
#12538