Michael McCandless wrote:
Mark Miller wrote:
Mark Miller wrote:
Michael McCandless (JIRA) wrote:
Well, it's tricky... and indeed all tests pass with the bug (which is
spooky -- I think we need to add cases to TestSort where 1) the index
has many segments, and 2) the number of hits is much greater than the
queue size), but I'm pretty sure it's a bug
Its a nasty one - I had a test I was running with 200,000 docs and
10 or 1000 length queue, match all docs query - seemed to come back
in order every time (though that doesnt mean its the right top n,
just that the top n returned were always in order). I'll try to
tweak that to catch this somehow.
Ah - I was getting very lucky, I guess over and over, by hand. As
soon as I loop it 10 times, one pretty much always fails.
OK, phew! But, then, does the test pass w/ the latest patch (which
should fix this issue)?
Mike
I whittled it down so that it always breaks on the first pass. The
values.length and numSlotsFull methods pass no problem, and the
incorrect method always failing now.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
