Skip to content

ci: fix three checks broken by new runner images - #2583

Open
Justinwold88 wants to merge 8 commits into
ruvnet:mainfrom
Justinwold88:contrib/ci-runner-image-fixes
Open

Justinwold88 wants to merge 8 commits into
ruvnet:mainfrom
Justinwold88:contrib/ci-runner-image-fixes

Conversation

@Justinwold88

@Justinwold88 Justinwold88 commented Nov 19, 2026 •

Copy link
Copy Markdown

Repair cold-runner bootstrap failures without changing sensing code or weakening tests. Current head 356fafeec5906830659bb400c4c209d84a87f742, base main; includes shared security baseline #2585 and preserves contributor history.

  • MQTT: digest-pinned non-root Mosquitto, read-only configuration, loopback exposure, actual publish/host reachability and owned cleanup. Standalone main retains the full MQTT-feature library suite and broker integration tests.
  • QEMU: checksum-verified Espressif QEMU package from the same digest-pinned ESP-IDF image in builder/consumers; remove floating source/bootstrap and stale fallback. Seven-case matrix/swarm/fuzz budgets retained; hosted Clang18 avoids unreliable APT setup.
  • Windows: enable long paths before Cargo checks out ruv-FANN.
  • Security:36.0.21 compatible patch and direct pinned native audit; upstream corrected its ranges, so no custom reconciliation/API calls remain.

Local actual broker execution, workflow syntax/ShellCheck, matrices/budgets, secret/diff checks passed. Previous heada93 terminal CI:40pass/5conditional skips, including full workspace, MQTT, QEMU/fuzz, Windows/macOS kernel, security and bench. The new change only simplifies the shared security workflow/docs; Rust sources are identical. Fresh current-head qualification is pending.

Landing: #2585 → this PR → #2545. Held #2513 incorporates this ancestry and preserves its focused mqtt:: selection because its long learner tests already run in full CI/Release. Preserve that selection when reconciling the lanes. No device operation, main merge, image publication, release or helper installation.

Run the MQTT test broker from the eclipse-mosquitto:2.0.18 image, build
Espressif QEMU on ubuntu-24.04, and enable git long paths before cargo
on Windows. Plan item 0.6.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T8hca9bTB4puaeBstXR9CE
Copilot AI balanced review requested due to automatic review settings November 19, 2026 11:52

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Changes recommended

The review evidence flags a moderate, workflow-breaking ordering issue where the Windows long-path workaround runs after rust-cache triggers cargo metadata, allowing a cold-cache run to hit the same path-too-long failure before the fix is applied.

1 open issue
What changed in this PR

Fixes three CI workflows that broke on newer GitHub-hosted runner images and in cold-cache scenarios.

Changes:

  • Run MQTT integration tests against a pinned eclipse-mosquitto:2.0.18 container instead of installing distro Mosquitto.
  • Pin the Espressif QEMU build job to ubuntu-24.04 to avoid GCC 15 build failures on newer runners.
  • Configure Git long-path support on Windows to prevent Cargo git dependency checkout failures.
File Description
.github/​workflows/​ruview-kernel-npm.yml Updates the Windows kernel CI workflow to enable Git long-path support to avoid path-length failures during Cargo git dependency operations.
.github/​workflows/​mqtt-integration.yml Switches MQTT integration tests to use a pinned Mosquitto Docker container with a mounted CI config and updated readiness/log steps.
.github/​workflows/​firmware-qemu.yml Pins the QEMU build job runner to ubuntu-24.04 to avoid GCC 15 warnings-as-errors in the bundled dtc build on newer images.

Review effort: Balanced
Copilot was unable to complete this review at the selected tier due to model policy or plan restrictions.

Comment on lines +83 to +85
- name: Allow long paths for cargo git checkouts
if: runner.os == 'Windows'
run: git config --global core.longpaths true
@proffesor-for-testing

Copy link
Copy Markdown
Collaborator

Taking the scoped runner repair/review under Dragan's explicit request; contributor ancestry and Windows long-path fix will be preserved. Current head25d867fb45944d3c85a3291ff1c75a5da0a19301 / main13b47883.

I am checking the cold-cache paths against #2513's already-qualified runner corrections. Planned follow-up on an isolated branch: reuse its digest-pinned SDK QEMU/runtime and non-root broker, preserve this PR's Windows fix, validate workflow syntax and actual affected jobs. The existing QEMU cache key otherwise still accepts old floating-source builds, and the broker tag is mutable. This copies only CI bootstrap changes, not the held server feature stack. No main merge, release or firmware/device action. Will post exact diff/head and results here before treating it ready.

Owner Dragan; agent Codex; host DraganMac; session ruview-server-repairs-20261119; lease until14:25UTC. Please flag conflicting edits to the three workflow files.

@proffesor-for-testing

proffesor-for-testing commented Nov 19, 2026 •

Copy link
Copy Markdown
Collaborator

Updated security assessment: raw CI still fails five entries, but primary-advisory checks refute four matches against core wasmtime36.0.20. Do not infer that a major Wasmtime migration is required.

Executed cargo-audit0.22.2 output from RustSec DB7133866a6b44d6d3e1994ae6a48a296553c3c3da is retained. Separate current upstream advisory HTML and repository security-advisory API readbacks show:

RustSec entry Primary advisory Applicability to this lockfile
0567 GHSA-2674-w29h-wqvj: affected branches start at39 Core36.0.20 is outside those ranges
0573 GHSA-wvgq-g5hf-6gm6: affected branches start at47 Core36.0.20 is outside those ranges
0571 GHSA-f8mm-ggr3-89wh: package wasmtime-wasi-http That package is absent; RustSec incorrectly matched core wasmtime
0566 GHSA-5m8g-xhq4-4h93: package wasmtime-wasi That package is absent; RustSec incorrectly matched core wasmtime
0569 GHSA-hj27-mpw9-ff4c: core versions below36.0.21 Version-applicable; a36.0.21 patch exists

STATIC caller inspection: Wasmtime is default-off in homecore-plugins/homecore-server. The current host uses default Config plus fuel and Caller::get_export; no direct Caller::get_module_export call was found. Registry36.0.20 source has distinct get_export/get_module_export implementations. This is bounded source evidence, not a complete runtime reachability or shipped-binary proof.

The smallest next remediation is a reviewed36.x patch/caller qualification plus correcting the four advisory package/range mismatches. No ignore, audit-threshold change, advisory database edit or major dependency migration has been introduced here. The raw audit remains failed and overall merge readiness stays held until that baseline is reconciled. The first five-entry report described scanner output; this primary-source assessment supersedes any inference that all five are actual defects in core36.0.20.

@proffesor-for-testing

Copy link
Copy Markdown
Collaborator

NO BLOCK in the reviewed source-repair scope at a93b78f04248ac53347ff0e9199cbbb418b1d53a. Full merge qualification remains pending; this is not maintainer approval or permission to merge.

Checked all changed runner workflows, same SDK digest in builder/consumers, removed stale floating-source QEMU cache, unchanged7-case QEMU matrix/fuzz budgets, pinned non-root loopback broker and cleanup, Windows long-path ordering. Real local broker startup and current-head hosted MQTT/QEMU bootstrap/fuzz/Windows kernel/audit/Semgrep pass. Remaining QEMU cases/macOS/full workspace/bench are not relabeled as passed.

Current checkpoint: {'IN_PROGRESS': 6, 'SUCCESS': 31, 'SKIPPED': 2, 'QUEUED': 2}; no failed checks observed. Still pending: QEMU Test (full-adr060), QEMU Test (edge-tier0), QEMU Test (tdm-3node), QEMU Test (boundary-min), Security Report, Rust Workspace Tests, kernel (macos-latest), bench compile-verify (--no-run). Source/targeted results above do not turn those jobs into executed passes. Shared baseline #2585 is included in the three contributor branches; all original ancestry is preserved through normal merges.

No main merge, helper/app replacement, plugin enablement, firmware flash or physical/customer acceptance is claimed. Reverify if head/base changes. Full current-head CI and rUv's separate review/merge decision remain the next gates.

@proffesor-for-testing

Copy link
Copy Markdown
Collaborator

NO BLOCK in the previously reviewed repair scope; final exact-head hosted qualification is now complete.

Terminal hosted qualification — 2026-11-20

Exact head a93b78f04248ac53347ff0e9199cbbb418b1d53a: 40 passed / 5 conditional skips; no failed, queued or running checks. Full CI, security and benchmark compilation completed successfully. Conditional skips remain skips. This closes the earlier pending-CI checkpoint; it is not maintainer approval, server merge, installed-helper qualification or hardware acceptance. Security retains the unmodified native report separately from the narrow live-primary-verified metadata reconciliation.

@proffesor-for-testing

Copy link
Copy Markdown
Collaborator

NO BLOCK in the reviewed Nov20 reconciliation delta at 356fafeec5906830659bb400c4c209d84a87f742; full current-head merge qualification remains pending.

Only shared security workflow/docs change relative to the previously qualifieda93 head; pinned broker/QEMU matrices, fuzz budgets and Windows long-path ordering preserved. Native upstream audit replaces temporary reconciliation. Fresh full qualification pending.

Earlier terminal CI receipts remain historical and are not relabeled as passes on this new head. Same-head duplicate push QEMU/bench and unfinished supersededb109 runs were cancelled only where an active replacement exists; completed failure evidence is retained. No main merge, installed-helper/app replacement, release or sensor operation. Maintainer review/exact-head merge authority remains separate.

This branch has not been deployed

No deployments
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.

4 participants