This is obviously a large constant factor of overhead, and we want to reduce it. In the meantime, we’ve changed the default history size to only 2000 revs and added a flag to change that number.
kr
Doozer shares as much of the data structure as possible between
revisions, so the overhead should be about the same regardless
of how many files exist in total.
> Is there an equation to estimate the memory
> usage as the number of entries/revisions grows?
Not that I know of. The memory use is going to change significantly
anyway, as we optimize.
kr