Skip to content

ci: robust CI workflow + fix --userdataobj flag#63

Closed
simonCatBot wants to merge 1 commit into
KhronosGroup:openvx_1.3from
simonCatBot:fix/ci-improvements-rustvx
Closed

ci: robust CI workflow + fix --userdataobj flag#63
simonCatBot wants to merge 1 commit into
KhronosGroup:openvx_1.3from
simonCatBot:fix/ci-improvements-rustvx

Conversation

@simonCatBot

Copy link
Copy Markdown
Contributor

This PR replaces #62 and fixes all CI failures discovered during the first run.

What is fixed

  1. Build.py — added missing --userdataobj CLI flag (same as fix(Build.py): add missing --userdataobj CLI flag #62)

  2. CI workflow robustness:

    • build-main now probes for --userdataobj support before using it, avoiding chicken-and-egg failures when the base branch does not yet have the flag.
    • build-debug preserves build/Linux/x64/Debug/ artifacts so downstream lcov coverage collection can find .gcno/.gcda files.
    • cts-baseline restored the lcov coverage step.
    • cts-neural-networks and cts-graph-features already marked continue-on-error for known pre-existing failures.

Verification

  • build-debug ✅
  • build-release ✅
  • build-cts ✅
  • Most CTS stages pass ✅
  • Known failures: AlexNet weights missing (neural networks), pipelining incomplete in C model (graph features)

… benchmarking (KhronosGroup#61)

Modeled after kiritigowda/rustVX conformance.yml workflow.

Phase 1 (parallel):
- build-debug: Debug build with all extensions + coverage flags
- build-release: Release build with all extensions (for benchmarking)
- build-cts: Build Khronos CTS against debug install

Phase 2 (parallel, needs build-cts):
- cts-baseline: Smoke tests + code coverage collection
- cts-enhanced-vision, cts-neural-networks, cts-ix,
  cts-graph-features, cts-data-objects, cts-user-kernels:
  Stage-based CTS runs with --filter, matching rustVX pattern

Phase 3 (PR only, needs build-release + build-main):
- build-main: Build target branch release for comparison
- perf-gate: Clone openvx-mark, build against PR and main,
  run benchmarks, compare results

Known issues (marked continue-on-error):
- cts-neural-networks: TensorNetworks.AlexNetTestNetwork fails
  to load weights (missing test_data file, pre-existing)
- cts-graph-features: GraphPipeline.* fails (pipelining incomplete
  in C model target, pre-existing)

Removed legacy .travis.yml and .gitlab-ci.yml.
Excluded --conf_nnef due to NNEF-Tools GCC 11 incompatibility.

Co-authored-by: Simon <simon@openclaw.ai>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants