refresh_bridges returns ok for up-to-date, not-on-tier, and failed alike; say which

by adom

Adom Bridge unleashes your AI onto your full PC: full power, full safety. The AI breaks out of the container it runs in and onto your real machine, managing and running your entire laptop as you: launch apps, move files, drive any window, control KiCad, Fusion 360 or your real signed-in browser. Works with any AI, cloud or local (Claude Desktop, Claude Code, Codex), no Hydrogen required. Bridge and its bridges are an operating system for AI, with managed Node and Python runtimes, and a human-onl

Overview Files Skills Discover Packages Releases Dependencies Issues19 Breadcrumbs Pull Requests1 Activity

Closed general

John Lauer · 2mo ago ·closed by John Lauer

Field report from adom/kicad-bridge#71 (Ray's agent on ROG-Zephyrus, verbatim there): the box runs kicad-bridge 0.9.272, insiders builds 0.9.273+ exist, and three refresh_bridges calls returned status: ok while bridgeVersion stayed 0.9.272.

A caller cannot tell these apart, because all three return ok:

you are up to date
a newer build exists but this login/box is not on the tier that carries it (or the installed bridge.json predates insidersManifestUrl, so the tier is never probed)
the update was attempted and failed

Ask: have refresh_bridges (and the updater's per-bridge report) say which of the three happened, per bridge: installed, available (per tier it probed), tierProbed (public / insiders / not declared), and outcome (up_to_date | newer_on_other_tier | updated | failed: <why>). The one the user can act on is the tier, so name it.

Same lesson as kicad-bridge's rescan_libraries: report what was achieved, not what was attempted. Not urgent; the KiCad-side workaround is the one-time bridge_install {"manifestUrl": <insiders version.json>, "force": true} bootstrap, which I have given Ray.

2 Replies

John Lauer · 2mo ago

Shipped in ab 2.1.11 (insiders; pkg published). refresh_bridges (single-name and sync-all) now merges three fields into every result row alongside action:

tierProbed: insiders | public | notDeclared — notDeclared is the actionable one you named: the installed bridge.json predates insidersManifestUrl (and no insiders install URL is persisted), so the tier is never probed no matter how many times you refresh. Ray's ROG-Zephyrus case reads as tierProbed:"notDeclared" now instead of a bare ok.
available: the version the probed tier's manifest offered (null when the fetch failed).
outcome: up_to_date | updated | failed: <why> (manifest not served at the probed tier / unparseable / wiki unreachable — each named).

One honest limitation, by design: a public-tier box cannot report "newer exists on insiders" because the insiders manifest is unreadable to it — that invisibility IS the tier mechanism. What it CAN and now DOES tell you is when it never probed the tier at all (notDeclared), which is the case a human can fix with the one-time bridge_install {"manifestUrl": <insiders version.json>, "force": true} bootstrap you gave Ray.

John Lauer · 2mo ago

Delivered in 2.1.11: refresh_bridges results carry tierProbed (insiders | public | notDeclared - the actionable case), available, and outcome (up_to_date | updated | failed: ) per bridge, for the single-name and sync-all forms. Ray's case reads as tierProbed notDeclared instead of a bare ok. The one honest limit stands: a public box cannot see the insiders tier's version - that invisibility IS the tier. Closing.

John Lauer closed this · 2mo ago