From umami's 0.10.0 dogfood (concern #3).
doctor --setup's sign-in failure path takes the adapter's full waitForURL timeout (~30s in umami) before reporting ✗. Correct, but it feels like a hang.
PR #26 added a progress line (… running adapter.setup (bounded by your adapter's own timeouts)) as a partial fix. This issue tracks the deeper improvement:
- The probe runs the adapter's real
setup, so tutorial-forge can't directly shorten the adapter's internal waits.
- Idea: an overridable probe deadline (
--setup-timeout <ms>) that races probeAdapterSetup and reports "adapter.setup didn't finish within Ns" — turning a 30s wait into a fast, clear signal.
Low priority; the progress line covers the worst of the UX.
From umami's 0.10.0 dogfood (concern #3).
doctor --setup's sign-in failure path takes the adapter's fullwaitForURLtimeout (~30s in umami) before reporting✗. Correct, but it feels like a hang.PR #26 added a progress line (
… running adapter.setup (bounded by your adapter's own timeouts)) as a partial fix. This issue tracks the deeper improvement:setup, so tutorial-forge can't directly shorten the adapter's internal waits.--setup-timeout <ms>) that racesprobeAdapterSetupand reports "adapter.setup didn't finish within Ns" — turning a 30s wait into a fast, clear signal.Low priority; the progress line covers the worst of the UX.