render: include olm.bundle.object properties in rendered bundle images

[joelanford]

Signed-off-by: Joe Lanford [email protected]

Description of the change:
In opm render, include olm.bundle.object properties in rendered bundle images.

Motivation for the change:
By default, we need bundles to include olm.bundle.object properties so that the GRPC server can serve
them to OLM's packageserver component, which is used by OLM clients (e.g. the kubectl-operator plugin) to provide user-friendly installation workflows and display information about the operators available in catalogs installed in the cluster.

Note: packageserver only actually uses the CSVs from the heads of each channel. Catalog maintainers may opt to delete objects from non-channel heads to reduce the size of the catalog.

Reviewer Checklist

• Implementation matches the proposed design, or proposal is updated to match implementation
• Sufficient unit test coverage
• Sufficient end-to-end test coverage
• Docs updated or added to /docs
• Commit messages sensible and descriptive

[openshift-ci]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: joelanford

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [joelanford]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[jianzhangbjz]

Thanks for your quick action!

Catalog maintainers may opt to delete objects from non-channel heads to reduce the size of the catalog.

Yes, this is a potential issue, but why not add a command or integrate some code to address this issue at this time? Or am I missing something? Thanks!

[codecov]

Codecov Report

Merging #807 (c72969f) into master (d5522b3) will decrease coverage by 0.15%.
The diff coverage is 68.96%.

❗ Current head c72969f differs from pull request most recent head 5b19561. Consider uploading reports for the commit 5b19561 to get more accurate results

@@            Coverage Diff             @@
##           master     #807      +/-   ##
==========================================
- Coverage   51.08%   50.92%   -0.16%     
==========================================
  Files         103      103              
  Lines        9056     8995      -61     
==========================================
- Hits         4626     4581      -45     
+ Misses       3553     3551       -2     
+ Partials      877      863      -14     

Impacted Files	Coverage Δ
pkg/registry/registry_to_model.go	59.72% <65.38%> (+0.02%)	⬆️
alpha/action/render.go	64.06% <100.00%> (+0.18%)	⬆️
pkg/registry/csv.go	71.42% <0.00%> (-5.45%)	⬇️
pkg/registry/query.go	60.41% <0.00%> (-0.53%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d5522b3...5b19561. Read the comment docs.

[joelanford]

Yes, this is a potential issue, but why not add a command or integrate some code to address this issue at this time? Or am I missing something? Thanks!

This is a completely reasonable thing to do, and you aren't missing anything.

We're methodically designing and building out the pieces that catalog maintainers will likely need, and we want to make sure we build the right abstractions. This "prune unnecessary bundle objects" feature likely falls in the broader category of catalog post-processing and optimizations, so we're holding off on integrating specific post-processors until we have a more holistic idea of how all the catalog pipeline pieces fit together.

[joelanford]

This is the breaking change that go-apidiff is complaining about. This function was introduced with FBC code, so it hasn't been in the wild long.

I'm perfectly willing to refactor to avoid a breaking change if others think its necessary here.

[kevinrizza]

/lgtm