refactor: decompose actions into testable modules#14
Merged
Conversation
…t config Add createActionContext() to build an ActionContext from pre-constructed dependencies, enabling tests to inject mocks without touching @actions/core. Add root vitest.config.ts and an nx test target so `npx nx affected --target=test` works across all workspaces. https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
Run `npx nx affected --target=test` after building. Zero-cost until workspaces add a test script — actions without one are simply skipped. https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
…rativeModel Add @gemini-actions/shared/testing entry point with: - createMockOctokit(): proxy-based stub for GitHub API calls - createMockModel(): canned-response GenerativeModel mock - createTestContext(): convenience wrapper combining both Exported separately so production code never imports test helpers. https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
Decompose monolithic index.ts into: - run.ts: runPrReview(ctx, inputs) orchestration function - prompts.ts: STRICTNESS_PROMPTS, buildFileSections, buildReviewPrompt - review.ts: filterValidComments, buildSummaryBody - index.ts: thin shell that reads inputs and delegates Add tests for all three modules (14 tests). https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
…tability
Decompose monolithic index.ts into:
- run.ts: runTestFailureDiagnosis(ctx, {prNumber, testOutput}) — receives
file content as string, eliminating the fs dependency from testable code
- parsers.ts: extractTestFilePaths() pure function
- index.ts: thin shell that reads the file and delegates
Add parser tests covering Jest, pytest, Go, generic patterns (9 tests).
https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
Decompose monolithic index.ts into:
- run.ts: runPrFromIssue(ctx, {issueNumber}) orchestration
- prompts.ts: buildPlanPrompt, buildChangePrompt, buildPrBody
- index.ts: thin shell
Add prompt tests (4 tests).
https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
…testability Decompose monolithic index.ts into: - datadog.ts: DatadogClient interface + createDatadogClient() factory, making the HTTP layer injectable for tests - run.ts: runDatadogResponder(ctx, inputs, ddClient) orchestration - prompts.ts: buildAnalysisPrompt() - index.ts: thin shell Add prompt tests (5 tests). https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
Decompose monolithic index.ts into: - globs.ts: matchGlob, matchesAnyGlob pure functions - prompts.ts: buildFileSelectionPrompt, buildAnswerPrompt - run.ts: runRepoQA(ctx, inputs) orchestration - index.ts: thin shell Add glob matching tests (9 tests). https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
…lity Decompose monolithic index.ts into: - registry.ts: RegistryResolver type + createRegistryResolver(fetchFn) with injectable fetch for testing - run.ts: runDependencyImpact(ctx, inputs, resolver) orchestration - index.ts: thin shell Existing parsers.ts, prompts.ts, review.ts, types.ts modules and their tests are unchanged. https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5
dortort
force-pushed
the
claude/refactor-actions-testing-QJgnS
branch
from
a1fd41c to
656d93f
Compare
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a1fd41c577
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+190
to
+191
| const parsed = parseJsonResponse<Step2Result>(step2Response); | ||
| step2Result = parsed; |
There was a problem hiding this comment.
parseJsonResponse only guarantees syntactic JSON (see shared/src/gemini.ts), so Gemini can return objects missing impacts/unaffectedUsages without throwing. In this block, assigning step2Result = parsed immediately means malformed Step 2 output can silently replace the safe { impacts: [], unaffectedUsages: [] } fallback, and Step 3 then synthesizes risk from incomplete/incorrect structure instead of the intended empty-impact baseline. Please keep a shape check (or access required fields) before assignment so bad Step 2 payloads fall back cleanly.
Useful? React with 👍 / 👎.
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
2 participants
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.
Extract each action's monolithic index.ts into a 3-layer architecture:
Shared infrastructure changes:
Per-action decomposition:
All 117 tests pass across 6 action workspaces.
https://claude.ai/code/session_01LhT9QaGdJ2J37cnBEULtL5