Cached names from web services may ignore some homonyms

[Thomas Kluyver]

e.g. Querying CoL, if we request //Glycyrrhiza glabra// first, //G. glandulifera// is cached as a synonym. If we then ask for //G. glandulifera//, it finds it in the cache, and doesn't look it up from the web service, so it misses that it's also a synonym of //G. uralensis//.

The cache should be structured so that we don't assume we know everything about a name we've only stored via synonymy. On the other hand, the extra HTTP requests would mostly be unnecessary, so this is low priority.