[st3penta]

What:

Set Codecov project and patch targets to 100% with zero tolerance.
Disable PR comments, fail CI when uploads fail, and remove the local
Makefile coverage threshold check.

Why:

Keep coverage targets in Codecov and surface upload failures in
pre-merge CI.

Supersedes: #1840
Ref: https://issues.redhat.com/browse/KFLUXDP-1172
Co-Authored-By: gpt-6-luna noreply@openai.com

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Enterprise
• Run ID: e77b7e96-73c7-4ad7-a610-2035946f3c96

📥 Commits

Reviewing files that changed from the base of the PR and between ed1b0dd and cdb53c4.

📒 Files selected for processing (3)

• .github/workflows/pre-merge-ci.yaml
• Makefile
• codecov.yml

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

---

📝 Walkthrough

Walkthrough

The changes update Codecov coverage targets and comment settings, remove the Makefile coverage threshold check, and make Codecov upload errors fail the pre-merge CI job.

Changes

Coverage CI configuration

Layer / File(s)	Summary
Coverage settings and CI behavior codecov.yml, Makefile, .github/workflows/pre-merge-ci.yaml	Codecov project and patch targets are set to 100% with 0% thresholds, and Codecov comments are disabled. The Makefile no longer checks for coverage below 100%. The Codecov upload step now fails CI if the upload reports an error.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Other

Suggested reviewers: robnester-rh, dheerajodha

Merge Risk: ⚪ Minimal · up to cdb53

The Codecov targets and upload-error handling match the stated intent. No material merge-blocking issue was identified.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title clearly and concisely describes the main change: setting Codecov coverage targets to 100%.
Description check	✅ Passed	The description includes the What and Why sections and provides the related Jira reference. It is complete enough to explain the change and its context, although the reference appears outside a dedica…

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

---

_{Comment @coderabbitai help to get the list of available commands.}

[fullsend-ai-review]

Risk Assessment: moderate (2/5)

Details

Small CI-configuration-only PR touching a CI workflow (primary risk driver) and Makefile/codecov config; no security-sensitive files, no dependency changes, known contributor; moderate composite risk.

[fullsend-ai-review]

No findings at or above the configured severity threshold.

[jsmid1]

lgtm

[Acepresso]

LGTM

[fullsend-ai-retro]

Retro for conforma/policy#1868

Workflow went well. This was a 31-line CI/config PR (new codecov.yml, Makefile 100% coverage check removed, fail_ci_if_error: true), human-authored (co-authored by an OpenAI Codex session). The review agent (fullsend-ai-review) correctly produced "no findings at or above the configured severity threshold"; CodeRabbit also had no actionable comments; two human contributors approved with bare "lgtm" / "LGTM". Merged cleanly 7 days after opening.

Autonomy-readiness note. The review agent's verdict matched both human reviewers' uncommented approvals on a trivial config-only PR. This is corroborating evidence for the already-open fullsend-ai/fullsend#8135 (file-type-based routing to a lighter review path for config-only PRs); the review here spent ~11:44 and spawned 6 sub-agents (correctness, security, intent-coherence, style-conventions, docs-currency, risk-assessment) for a 3-file diff. Not re-filing per retro policy — logging the evidence here instead.

Filing one proposal: a defect in the platform's NFKC sanitizer telemetry that made the sub-agent transcripts for this run appear to contain tens of thousands of fullwidth characters when there were almost certainly only a handful of compatibility code points. The underlying workflow was unaffected (nothing blocked, success), but the misleading metric wasted my own analysis bandwidth during this retro and would mislead any future investigator.

Other things noticed, not filed. PR #1840 is still open despite PR #1868's description saying it supersedes it — one-off cleanup, not systemic. Two duplicate pull_request_target dispatches landed 10s apart at 2026-10-07 15:30 (both correctly skipped by downstream dedup) — only one occurrence, insufficient signal to propose. Historical retro runs in conforma/.fullsend have been failing since 2026-09-16; this retro successfully ran in conforma/policy instead, so the migration appears to be working — no action needed.

Duplicate-check performed against conforma/policy, conforma/.fullsend, fullsend-ai/fullsend, and fullsend-ai/agents. The sanitizer-metric proposal is novel; no overlap with #7391, #8135, #4183, #2275, #2383, #2646, agents#397, agents#230, agents#693, agents#394, agents#435, or agents#150.

Proposals filed

• NFKC sanitizer's "characters affected" count overreports by orders of magnitude (in fullsend-ai/fullsend)