Fix Codex Minimal Reasoning Effort Execution#2020
Open
QwertyCoolMT wants to merge 1 commit into
Open
Conversation
📝 WalkthroughWalkthroughExtends ChangesCodex Workflow-Level Options and Minimal Effort
Test Environment Variable Isolation
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
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 |
- Forward workflow-level Codex options into DAG assistant config - Preserve explicit reasoning effort precedence over preset tiers - Add regression and validation coverage for minimal effort
QwertyCoolMT
force-pushed
the
archon/task-fix-archon-minimal-setting-retry2
branch
from
9b7d934 to
6853b05
Compare
QwertyCoolMT
changed the title
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 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 `@packages/server/src/routes/api.workflows.test.ts`:
- Around line 16-31: The `originalArchonHome` variable is captured once at
module load time, which means `afterEach` restores a potentially stale value
instead of the test's actual incoming value. Move the `originalArchonHome`
assignment from module scope into the `beforeEach` hook as a `let` variable
instead of `const`, so that each test run captures its own incoming state of
`process.env.ARCHON_HOME` before modifying it. This ensures that `afterEach`
restores the environment to the correct pre-test state for each individual test,
maintaining proper test isolation.
🪄 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: defaults
Review profile: CHILL
Plan: Pro
Run ID: 3e08bde2-9b16-4405-a3e4-90a97c1af1da
📒 Files selected for processing (9)
packages/adapters/src/forge/github/adapter.test.tspackages/cli/src/commands/ai.test.tspackages/core/src/db/codebases.test.tspackages/providers/src/codex/provider.test.tspackages/server/src/routes/api.providers.test.tspackages/server/src/routes/api.user-ai-prefs.test.tspackages/server/src/routes/api.workflows.test.tspackages/workflows/src/dag-executor.test.tspackages/workflows/src/dag-executor.ts
Comment on lines
+16
to
+31
| const originalArchonHome = process.env.ARCHON_HOME; | ||
| const isolatedArchonHome = join(tmpdir(), 'archon-api-workflows-test-home'); | ||
|
|
||
| beforeEach(async () => { | ||
| await rm(isolatedArchonHome, { recursive: true, force: true }); | ||
| process.env.ARCHON_HOME = isolatedArchonHome; | ||
| }); | ||
|
|
||
| afterEach(async () => { | ||
| await rm(isolatedArchonHome, { recursive: true, force: true }); | ||
| if (originalArchonHome === undefined) { | ||
| delete process.env.ARCHON_HOME; | ||
| } else { | ||
| process.env.ARCHON_HOME = originalArchonHome; | ||
| } | ||
| }); |
There was a problem hiding this comment.
Make ARCHON_HOME snapshot per-test, not module-scoped.
originalArchonHome is captured once at module load (Line 16), so afterEach restores a potentially stale value instead of the test’s incoming value. Capture it inside beforeEach via a mutable let to keep restoration symmetric per test run.
Suggested patch
-const originalArchonHome = process.env.ARCHON_HOME;
+let originalArchonHome: string | undefined;
const isolatedArchonHome = join(tmpdir(), 'archon-api-workflows-test-home');
beforeEach(async () => {
+ originalArchonHome = process.env.ARCHON_HOME;
await rm(isolatedArchonHome, { recursive: true, force: true });
process.env.ARCHON_HOME = isolatedArchonHome;
});As per coding guidelines, “Keep tests deterministic … [and ensure] test isolation.”
🤖 Prompt for 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.
In `@packages/server/src/routes/api.workflows.test.ts` around lines 16 - 31, The
`originalArchonHome` variable is captured once at module load time, which means
`afterEach` restores a potentially stale value instead of the test's actual
incoming value. Move the `originalArchonHome` assignment from module scope into
the `beforeEach` hook as a `let` variable instead of `const`, so that each test
run captures its own incoming state of `process.env.ARCHON_HOME` before
modifying it. This ensures that `afterEach` restores the environment to the
correct pre-test state for each individual test, maintaining proper test
isolation.
Source: Coding guidelines
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Describe this PR in 2-5 bullets:
modelReasoningEffort: minimalwere parsed but not forwarded into DAG nodeassistantConfig.minimalreasoning could get a different runtime effort than configured, especially when tier defaults or aliases were involved.modelReasoningEffort,webSearchMode, andadditionalDirectoriestoassistantConfig, with explicit workflow reasoning taking precedence over preset effort.effort, Copilot/OpenCode behavior, Settings UI semantics, docs, generated API types, and unrelated workflow YAML/defaults drift.UX Journey
Before
After
Architecture Diagram
Before
After
Connection inventory (list every module-to-module edge, mark changes):
workflowSchemaloaderloaderdag-executordag-executorassistantConfigdag-executormodelReasoningEffortblocks preset effort overwrite for Codex.assistantConfigCodexProvidereffort: minimal.Label Snapshot
risk: lowsize: Mworkflows|providers|server|cli|adapters|core|testsworkflows:dag-executor,providers:codex,server:ai-config,cli:ai,tests:validationChange Metadata
bugmultiLinked Issue
1d7c2db9d26534d28e79af44b0d707a6Validation Evidence (required)
Commands and result summary:
bun run type-check bun run lint --max-warnings 0 bun run format:check bun run test bun run build bun run validatebun run testpass with 5135 passed / 0 failed / 17 skipped, build pass, and aggregatebun run validatepass.Security Impact (required)
Yes/No): NoYes/No): NoYes/No): NoYes/No): NoYes, describe risk and mitigation: N/ACompatibility / Migration
Yes/No): YesYes/No): NoYes/No): NoHuman Verification (required)
What was personally validated beyond CI:
modelReasoningEffort: minimalreachesassistantConfig;webSearchModeandadditionalDirectoriesare forwarded for Codex; built-in Codexsmalltier still routes minimal effort; explicit workflow reasoning wins over preset effort.effort: minimal; per-user tier/alias writes accept Codexeffort: minimal; CLI--effort minimalvalidation accepts Codex.Side Effects / Blast Radius (required)
Rollback Plan (required)
6853b056.minimaleffort fail.Risks and Mitigations
codex.Summary by CodeRabbit
New Features
Tests