On Wed, Sep 5, 2012 at 11:26 AM, Dawid Weiss <[email protected]> wrote: >> But: it looks like it was happening due to RamUsageEstimator's >> overhead of tracking which objects it has seen ... so I committed a >> fix to do its own rough accounting. > > This can amount to some cost, true, the cost of the set (table of references).
Yeah ... this test in particular is object-heavy (one object per term doc+position occurrence!) so I think it amounted to non-trivial extra RAM ... But it was odd I could not repro ... Mike McCandless http://blog.mikemccandless.com --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
