Skip to content

fix(app): keep model controls available during generation - #2958

Merged
reachjalil merged 1 commit into
different-ai:devfrom
reachjalil:feature/enable-selectors-during-load
Jul 21, 2026
Merged

fix(app): keep model controls available during generation#2958
reachjalil merged 1 commit into
different-ai:devfrom
reachjalil:feature/enable-selectors-during-load

Conversation

@reachjalil

Copy link
Copy Markdown
Collaborator

Summary

  • keep the composer model and thinking-mode selectors available during ordinary assistant generation
  • track steered runs separately and disable both selectors from steer submission until the run returns idle
  • close/guard the compact model picker while steering
  • add regression coverage for the normal-generation versus steering distinction

Why

The composer previously disabled model and thinking selection for every busy session. That prevented users from preparing the next run while an ordinary response was loading. Steering remains protected because changing execution settings during an active steer can cause inconsistent behavior.

Checks

  • pnpm --filter @openwork/app exec bun test tests/composer-model-controls.test.ts — passed (2 tests)
  • pnpm --filter @openwork/app test — passed (372 tests)
  • pnpm --filter @openwork/app typecheck — passed
  • isolated Electron app-smoke eval on the Hub-assigned CDP port — passed

Manual verification

Not run — the isolated merge-readiness profile was fresh and had no onboarded workspace or usable model, so a real generation/steer journey was unavailable.

Reviewer steps:

  1. Start a task with a configured model.
  2. While the first response is generating, confirm model and thinking selectors remain usable.
  3. Submit a steering follow-up while that response is active.
  4. Confirm both selectors disable for the steered run and re-enable when the run returns idle.

Risk

Low and UI-local. No API, IPC, persistence, or model request shape changes.

@vercel

vercel Bot commented Jul 21, 2026

Copy link
Copy Markdown
Contributor

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
openwork-landing Ready Ready Preview, Comment, Open in v0 Jul 21, 2026 11:30am
1 Skipped Deployment
Project Deployment Actions Updated (UTC)
openwork-diagnostics Skipped Skipped Jul 21, 2026 11:30am

@vercel
vercel Bot temporarily deployed to Preview – openwork-diagnostics July 21, 2026 11:29 Inactive
@vercel

vercel Bot commented Jul 21, 2026

Copy link
Copy Markdown
Contributor

@reachjalil is attempting to deploy a commit to the Different AI Team on Vercel.

A member of the Team first needs to authorize it.

Copy link
Copy Markdown
Collaborator Author

@benjaminshafii This is ready and being merged into dev. The fix keeps model/thinking selection available during ordinary generation while preserving the steering lock.

@reachjalil
reachjalil merged commit 95c9ebb into different-ai:dev Jul 21, 2026
7 of 10 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant