The default HGAtomCache implementation is optimized for typical usage but less
so for bulk import of data. Furthermore, caching is also performed by the
storage layer and by the OS, so caching might be done redundantly, which might
also cause bigger memory consumption than strictly necessary.
HGAtomCache used should be defined in HGConfiguration, so it can be replaced by
custom, storage-engine-based lightweight caches or temporarily disabled by
using a fake cache for certain import scenarios.
Original issue reported on code.google.com by Ingvar.Bogdahn on 4 Aug 2013 at 10:11
Original issue reported on code.google.com by
Ingvar.Bogdahn
on 4 Aug 2013 at 10:11