Repository navigation
[infra] Clean up nightly release artifacts - #51
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Enterprise Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughThe nightly workflow validates ChangesNightly release pipeline
Estimated code review effort: 4 (Complex) | ~45 minutes Sequence Diagram(s)sequenceDiagram
participant CheckChanges
participant MatrixBuildJobs
participant NightlySummary
participant ReleaseJob
participant GitHubActionsArtifacts
CheckChanges->>MatrixBuildJobs: provide validated package_version
MatrixBuildJobs->>GitHubActionsArtifacts: upload versioned artifacts
NightlySummary->>GitHubActionsArtifacts: download matching staging artifacts
NightlySummary->>GitHubActionsArtifacts: upload consolidated evaluation record
ReleaseJob->>GitHubActionsArtifacts: download cuvslam-* release assets
ReleaseJob->>ReleaseJob: validate assets and extract release notes
Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/workflows/nightly.yml:
- Line 270: Replace every mutable version-tag reference for
actions/upload-artifact, actions/download-artifact, and actions/github-script in
the nightly workflow with the corresponding immutable commit SHA, including the
shown actions/upload-artifact@v7 usage. Preserve each action’s current version
and configuration while pinning all occurrences consistently.
In `@cuvslam-skills/cuvslam-ci/SKILL.md`:
- Line 60: Update the nightly eval-runner requirements in
cuvslam-skills/cuvslam-ci/SKILL.md at line 60 to require the RUNNER_STORAGE_ROOT
mount and configured repository secrets/variables, while stating that AWS CLI is
supplied by the cuvslam-ci:local CI image. In
cuvslam-skills/cuvslam-ci/reference.md at line 95, remove AWS CLI from
host-runner prerequisites and retain the detail that it executes from the CI
image.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Enterprise
Run ID: c9c7c743-c826-4533-abe6-cdf27c67b51e
📒 Files selected for processing (4)
.github/workflows/nightly.ymlcuvslam-skills/cuvslam-ci/SKILL.mdcuvslam-skills/cuvslam-ci/reference.mdscripts/package_cpp_dist.sh
Test Results
cuVSLAM Evaluation KPIs
Artifacts |
768d353 to
0e418ac
Compare
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/workflows/nightly.yml:
- Around line 662-665: Update the artifact-count validation around cpp_assets
and wheel_assets to derive expected counts from the artifact-producing matrix
configuration or a previously emitted count, rather than hardcoding 6. Ensure
the check remains aligned when matrix rows or skip_artifacts/wheel settings
change, while preserving the existing error and exit behavior.
- Around line 657-663: Normalize PACKAGE_VERSION to its PEP 440 form before
constructing the wheel_assets glob and applying the later wheel filename
substring validation, while preserving the existing C++ asset matching with the
raw version. Ensure suffixed versions such as 1.2.3-rc1 match wheel filenames
containing 1.2.3rc1.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Enterprise
Run ID: 45e9714e-9473-4677-82ba-dcb4f18f2949
📒 Files selected for processing (1)
.github/workflows/nightly.yml
…nvidia-isaac#51) * [fix] Fix USE_RERUN build breakages in euroc example and slam library Two regressions only reachable in USE_RERUN=ON builds, so the default (USE_RERUN=OFF) CI never caught them: - examples/euroc/cpp/track_euroc.cpp: the rerun trajectory path referenced an undeclared `slam_pose`. The Slam::Track()/GetPose() split removed the old variable; fetch the pose via slam.GetPose(). Also drop the now-unused <memory> include. - libs/slam/CMakeLists.txt: async_slam.cpp calls pipelines::logTrajectory() under USE_RERUN but slam did not link pipelines, breaking the slam_test link. Add pipelines to slam's USE_RERUN libraries (no dependency cycle).
Make nightly artifacts versioned and release-ready so the exact tested files are promoted unchanged with concise summaries.
Summary by CodeRabbit
New Features
Bug Fixes