Repository navigation
Conversation
…7241) The review on #17241 produced seven findings. All seven were valid: three were real bugs in code written that hour -- untagged staging tasks that --tags would skip, three GPU pip tasks reading manifests out of the controller's checkout, and a fixed-count scan that could stop inside a task result -- and one exposed a detector blind spot that turned a six-site defect class into sixteen. So the substance is not the noise, and profile: assertive stays. What made the review expensive to read was the per-comment wrapper: a restatement of each finding addressed to an AI agent, plus walkthrough chrome nobody here acts on. Off: enable_prompt_for_ai_agents (the largest block per comment, duplicating the prose above it), sequence_diagrams, estimate_code_review_effort, suggested_labels and suggested_reviewers (labels and reviewers follow the repo's own rules), and related_issues/related_prs (guessed from text, wrong often enough that checking each costs more than finding the real link by hand). Unchanged: profile, high_level_summary, review_status, collapse_walkthrough, path_filters and path_instructions. Not addressed: the 'Supported by static analysis' blocks, which are the other bulky part. The published schema exposes no key for them, and I would rather say so than switch off something adjacent and claim it worked.
|
Warning Review limit reachedNext included review available in 2 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Repository: mrveiss/AutoBot-AI/.coderabbit.yaml Review profile: ASSERTIVE Plan: Advanced Run ID: 📒 Files selected for processing (1)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
✅ SSOT Configuration Compliance: Passing🎉 No new hardcoded values of either class — Known backlog in |
|
Folded into #17241 ( |
1 participant
Thinking Path
The review on #17241 was hard to read. The instinct is to turn CodeRabbit down — but the
findings were the wrong thing to turn down. All seven were valid:
--tags backendrun would skip; three GPU pip tasks reading manifests out of the controller'scheckout; and an ignored-failure scan capped at a fixed line count that could stop inside a
task result
what they do with a path, so it reported a six-site defect class that is actually sixteen
A reviewer with that hit rate should not be made quieter. So
profile: assertivestays.What actually cost reading time was the wrapper around each finding, not the finding.
What Changed
One file,
.coderabbit.yaml,reviews:block only.enable_prompt_for_ai_agentsfalsesequence_diagramsfalseestimate_code_review_effortfalsesuggested_labelsfalsesuggested_reviewersfalserelated_issues,related_prsfalseUnchanged:
profile: assertive,high_level_summary,review_status,collapse_walkthrough,path_filters,path_instructions, and thebase_branchesoverride that keeps therelease-promotion PR reviewed.
Verification
Keys checked against the published schema (
coderabbit.ai/integrations/schema.v2.json) beforebeing written — each exists, is a boolean, and defaults to
true, so each line here is a realchange rather than a no-op. YAML parses; the
reviewsblock was re-read after the edit and thetwelve booleans are as intended.
What this does not fix, stated rather than implied: the "Supported by static analysis" blocks
are the other bulky part of each comment, and the schema exposes no key for them. I would rather
name that gap than switch off something adjacent and report the problem solved. If they remain
the main irritant, the next lever is a support request to CodeRabbit, not a config key.
Scope: deliberately separate from #17241. That PR is a deploy fix mid-incident; a reviewer
configuration change does not belong in its diff or its
Closesline.Model Used
Claude Opus 5 (1M context)
Refs #17241