feat(desktop): first-run inference-source onboarding#550
Open
LuigiElleBalotta wants to merge 23 commits into
Open
feat(desktop): first-run inference-source onboarding#550LuigiElleBalotta wants to merge 23 commits into
LuigiElleBalotta wants to merge 23 commits into
Conversation
On Windows, http://localhost connects via IPv6 first, finds nothing listening, and falls back to IPv4 after ~2s - racing health()'s 2.0s timeout and causing flaky 'unreachable' results for healthy local engines (LM Studio, vLLM, etc.). 127.0.0.1 connects directly.
OllamaEngine._DEFAULT_HOST changed to 127.0.0.1 in a prior commit; these respx mocks still targeted localhost, breaking 5 tests.
…h to 127.0.0.1 Final review for the localhost->127.0.0.1 fix found two leftover localhost defaults missed by the per-engine sweep: the abstract _OpenAICompatibleEngine._default_host (dead in practice since every concrete engine overrides it, but contradicted the updated docs) and the OllamaModelList fetch in SettingsPage.tsx, which hits the same Windows localhost/IPv6 DNS delay this whole change exists to avoid.
Prevents double setInferenceSource/startBackend calls via the onboarding form path while a submission is in flight.
Avoids a transient reappearance of the stale error banner between start_backend's server-side status reset and the next 800ms poll tick.
try_lock() on a freshly-constructed Mutex always succeeds today, but silently stranding status at phase: "starting" forever if it ever didn't would recreate the first-run dead end this feature fixes. Spawn an async task with .lock().await instead.
LuigiElleBalotta
requested review from
ANarayan,
jonsaadfalcon and
robbym-dev
as code owners
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.
Summary
~/.openjarvis/inference.jsonis missing, the app probes for a running Ollama/LM Studio instance, lets the user confirm or pick a custom OpenAI-compatible server, then boots the backend.InferenceSourceFormcomponent (used by Settings, Onboarding, and a new "Reconfigure" error-recovery flow onSetupScreen).SetupScreen's error state so a bad/unreachable inference source can be fixed without manually editinginference.json.config_missing,probe_local_engines(Ollama/LM Studio probes),start_backendstatus reset for clean retries.127.0.0.1(Windows localhost/IPv6 fix).Design doc:
docs/superpowers/specs/2026-06-12-inference-onboarding-design.mdImplementation plan:
docs/superpowers/plans/2026-06-12-inference-onboarding.mdTest plan
cargo test— 26/26 passcargo check— clean, no warningsnpx tsc -b --noEmit— no errorsnpm test— 6/6 passnpm run tauri devwith noinference.json— app builds and launches onboarding successfullyThis PR references #517 and it's made by Claude Code