Skip to content

fix(desktop): separate LAN and Tailscale pairing endpoints - #9882

Merged
maria-rcks merged 2 commits into
pingdotgg:mainfrom
maria-rcks:fix/triage-lan-endpoint
Sep 5, 2026
Merged

maria-rcks merged 2 commits into
pingdotgg:mainfrom
maria-rcks:fix/triage-lan-endpoint

fix(desktop): report only actual network exposure fallbacks

30a4991
Select commit
Loading
Failed to load commit list.
MacroscopeApp / Macroscope - Effect Service Conventions succeeded Sep 5, 2026 in 2m 24s

All clear

All clear

Details

Note

Your check run agent prompt is: .macroscope/check-run-agents/effect-service-conventions.md
More information about how Check Run Agents work can be found in our Help Center.


🔍 Agent activity — 3 tool calls

1. browse_code — apps/desktop/src/backend/DesktopServerExposure.ts

2. browse_code — apps/desktop/src/app/DesktopApp.ts

💭 I'm checking whether the named import of isTailscaleIpv4Address from the tailscale package is a pure helper, which fits the convention of allowing named imports from whole packages. I want to verify this function actually exists and is pure in the package source.

3. browse_code — In packages (or wherever @t3tools/tailscale lives), show the definition and export of isTailscaleIpv4Address and how it is exported from the package index. Is it a pure helper function or an Effect service? Was it added recently?

💭 I'm looking at the actual production changes: the DesktopApp condition tweak, the isUsableLanIpv4Address helper, and the unavailable-state check in resolveRuntimeState, confirming the import style is fine. Everything else follows existing conventions, but since backend behavior is changing, focused tests will be needed. Tests exist in DesktopServerExposure.test.ts. There's a minor inline duplication in resolveRuntimeState's network interface scan instead of a shared helper, but that's not an actual convention violation, and nothing else stands out under the error handling or service conventions. Since there are no real findings, I should report "All clear" -- I'll structure the complete_check call so the summary and title reflect that verdict cleanly.


Agent Credits: 43 credits

Token usage: 14 uncached input · 69,565 cache-read input · 45,461 cache-write input · 3,416 output
Agent Credits may also include non-token charges from external tools such as web research.