bug: circular-deps goroutine not cancelled when graph fetch fails in GetGraphAndCircularDeps

[claude]

Problem

In internal/api/client.go (GetGraphAndCircularDeps), two goroutines are launched concurrently — one for the supermodel graph and one for circular dependencies. If the graph request fails, the function returns early at line 568-570 without cancelling the circular-deps goroutine.

The circular-deps goroutine continues polling the API until the caller context is eventually cancelled, wasting API round-trips and holding connections open unnecessarily. Both channels are correctly buffered so there is no goroutine leak, but the goroutine does unnecessary work.

Fix

Introduce a derived cancellable context scoped to GetGraphAndCircularDeps, and cancel it via defer. This covers both the early-return and the happy-path cases:

inner, cancel := context.WithCancel(ctx)
defer cancel()
// pass inner to both pollJob calls instead of ctx

Location

• internal/api/client.go, lines 530-570 (GetGraphAndCircularDeps)

@claude please implement this

[claude]

Claude finished @claude[bot]'s task in 1m 4s —— View job

---

Result

The fix described in this issue is already implemented in the current codebase.

At internal/api/client.go lines 518–519, GetGraphAndCircularDeps already contains:

ctx, cancel := context.WithCancel(ctx)
defer cancel()

Both goroutines (lines 535 and 554) pass this derived ctx to pollJob, so:

• Early-return path (graph fetch fails, lines 568–570): defer cancel() fires and cancels the circular-deps goroutine
• Happy path: defer cancel() also fires on function return, cleaning up promptly

This matches the fix proposed in the issue exactly (using ctx shadowing rather than an inner variable, but with identical semantics). No code changes are needed.

[coderabbitai]

⚠️ Possible Duplicate Issue(s)

• GetCircularDependencies HTTP request not cancelled when GetGraph fails #78

🔗 Related PRs

#18 - feat: show circular dependency warning in context bomb [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!