You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[Bug]: Mobile home header stays on "Reconnecting" while one environment is asleep and another is connected #19074
I searched existing issues and did not find a duplicate.
I included enough detail to reproduce or investigate the problem.
Area
apps/mobile
Steps to reproduce
Pair the mobile app with two environments, for example an always-on machine and a laptop.
Let both sync at least one thread, then make one of them unreachable (put the laptop to sleep, or stop its server).
Open the mobile app while the other environment is connected, and look at the home header.
Leave it open, or come back hours later.
Expected behavior
With one environment connected the app is fully usable, so the header keeps the T3 Code brand. An environment that cannot be reached is reported as unreachable where it appears, not as a reconnect in progress.
Actual behavior
The header replaces the brand with "Reconnecting to " and stays that way for as long as the other machine is unreachable, which can be hours.
Two rules combine to cause it:
workspace-connection-status.ts shows the status whenever any enabled environment is connecting, even when another one is connected.
The connection supervisor retries forever with backoff up to 5 minutes, and both the backoff and the retry are presented as reconnecting (packages/client-runtime/src/connection/presentation.ts), so the label never changes to a failure.
With every environment down, the header says "Reconnecting 2 environments" indefinitely in the same way, although the attempts have already failed.
One environment connected, one unreachable
No environment reachable
In the first screenshot "homelab" is connected and its thread is live; "laptop" has been unreachable for hours.
Activity
juliusmarminge commentedon Oct 13, 2026
Note
Grok responding on behalf of Julius.
Open fix PR #19075 already covers this (same reporter). Added
Closes #19074there; no separate triage.