The persistent map to which the data is flushed or from which it is retrieved when not residing in cache.
The persistent map to which the data is flushed or from which it is retrieved when not residing in cache.
This method should be invoked from the implementations flush hook after it has determined the terminal version at which the entries in the cache are written to the persistent store.
This method should be invoked from the implementations flush hook after it has
determined the terminal version at which the entries in the cache are written
to the persistent store. If this method is not called, the cache will just
vanish and not be written out to the persistent store.
the new version to append to the paths in the cache (using the PathLike's addTerm method)
the current transaction (should be in commit or right-before commit phase)
Removes an entry from the cache, and only the cache.
Removes an entry from the cache, and only the cache. This will not affect any
values also persisted to persistent! If the cache does not contain an entry
at the given key, this method simply returns.
key at which the entry is stored
the current transaction
A cache map puts an in-memory transaction local cache in front of a persistent store. Entries written during the transaction are held in this cache for fast retrieval. But the cache serves a second purpose: In the confluent system, the write paths are incomplete during the transaction, as it is not known in advance whether a meld forces a new index tree to be generated or not. In this case, the implementation needs to gather this information during the transaction, and when the flush is performed, the new terminal version is appended before writing the cached entries to the persistent store.
the underlying system
the key type (typically
Intfor a variable map orLongfor an identifier map)