Documentation
¶
Overview ¶
Package registrykey adapts this onix deployment's own definition.RegistryLookup into a crawler.KeySource -- the generic key-lookup function catalog-core's Fetcher calls internally to verify a fetched artifact's signature. catalog-core knows nothing about onix's Subscription/status model, so the base64 decoding, usable-status check, and fault classification live here.
Shared by catalogcrawler and catalogpublisher: both need to resolve {nodeID, keyID} against the real registry and classify the result the same way (a down/rate-limited registry is transient and worth retrying; a missing/unusable/malformed key is permanent and never will be), so a future fix to that classification (a new model.Status value, a changed encoding convention) only has one place to land instead of drifting between two copies. Lives under pkg/plugin/implementation/internal specifically so plugin implementation packages still don't cross-import each other directly -- only this common, dependency-free helper is shared.
No caching is added at this layer: definition.RegistryLookup implementations (registry, dediregistry) already cache Lookup results themselves; a second, independently-expiring cache here would risk serving a revoked/rotated key past the registry plugin's own cache invalidation.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Resolve ¶
func Resolve(ctx context.Context, caller string, reg definition.RegistryLookup, nodeID, keyID string) (ed25519.PublicKey, error)
Resolve asks the registry for {subscriberID, keyID} and turns the answer into a usable Ed25519 key, or into a correctly classified failure.
func Source ¶
func Source(caller string, reg definition.RegistryLookup) crawler.KeySource
Source adapts reg into a crawler.KeySource. caller labels every error this produces (e.g. "catalogcrawler", "catalogpublisher") so a log line or PublishError/CrawlError reason still says which plugin hit it.
Types ¶
This section is empty.
Source Files