Since Sculptor version
The cache is useful when reading a file without a lock. And better when the remote file is not modified a lot (as the refresh check frequency can be reduced)
The interface to the service has been implemented only in the Sculptor interpreters sage[c,w,wc,wcd]/srep[c,w,wc,wcd]
If the cache data source is updated by other processes, the cache has to be discarded and reloaded. This process is called refresh. The refresh process can be cofigured with the ![o]file cache_auto_refresh attributte or manually triggered with the refresh_cache command.
There are two types of cache with regard to how the file cache is filled:
With an Access cache when a read operation in the cache fails, it tries to fetch it on the source, while on the All cache when a read fails, the read returns NSR.
The next, prev, find, match, count,… take profit of an All cache but not an Access cache.