[ZODB-Dev] Another possible ZODB optimization

[Jeremy Hylton]

>>>>> "TD" == Toby Dickenson <tdickenson@geminidataloggers.com> writes:

  TD> How does Jim's fsIndex compare at this?

Jim mentioned the other day that he thinks fsIndex will use more
memory when the keys are first inserted than it will when it reads the
index from the filesystem on the second try.  The btree buckets will
leave some wasted space as they grow for insertions of increasing
keys.  When the index is read from disk when the storage is opened, it
allocates exactly the right amount of space.

Jeremy