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
This commit was created on GitHub.com and signed with GitHub’s verified signature.
Device API format 11: the error code system_version_unverified and the error fields version and reason are gone, and a system's writes state in GET /v1/devices/{deviceId} is replaced by blockedActions (the actions this device's client cannot run yet; each answers device_outdated until the client updates). An action whose system runs a program version Deepdock has not yet approved for it answers the new 409 version_not_covered (CLI exit 2): nothing ran, retrying gets the same answer, and it works once Deepdock approves that version or the customer runs an approved one. The bundled TypeScript and Python clients, the skill's error reference and the device types follow. This needs the API from the Deepdock release that carries these changes.
deepdock devices and deepdock incidents list read the organisation's lists as { "data": [...], "nextCursor": ... }, the form /v1 already uses (was { "data": { "items": [...], "nextCursor": ... } }). They need the API from the next Deepdock release; an older API answers the old form, which this version does not read.
The CLI follows the API's British spelling and its renamed code: GET /partner/organisations (was /partner/orgs), the error code organisation_required (was organization_required) and cli_outdated (was client_outdated, still exit 4). The bundled device types and the skill's error reference include them. These need the API from the next Deepdock release.
The command group org is now orgs (deepdock orgs list, deepdock orgs select); the old name is gone, with no alias.
A device is named <deviceId> in every usage string, help text, the README and the bundled skill (devices rename|release|systems <deviceId>, actions list <deviceId> <system>).
The README and the bundled skill advise reusing one HTTP connection for calls (keep-alive), because a fresh connection can hit a cold start.
The docs, the bundled skill and the device API pin say "pair" for taking a device into your organisation (was "claim"); the rate_limited and pairing_code_invalid texts follow, and requestId is described as present on every error body. Field names and error codes are unchanged (claimedAt stays).