Slava Pestov on 2009-08-04 16:07:15 -0500:

> I noticed that sometimes, this test fails in bloom-filters on the
> various build machines, preventing binaries from uploading:

I've pushed a fix for the symptoms to git://github.com/alec/factor.git
branch bloom-filters-fewer-false-positive-test-failures.  Direct link:
http://github.com/alec/factor/commit/35701183f334e0bb9019bd0dc03b7b19fad9650a

The test asserts that the false positive rate is below a certain
threshold.  I ran it through 50,000 iterations and recorded the false
positive rates.  The maximum I saw was 16.8% (test was configured to
fail at 15%).  The patch bumps it up to 30% as overkill; that should be
enough to catch any usefulness-crippling changes but not be in any
danger of failing due to the randomness in the test.

> Did you plan on working on the vocabulary some more?

Plan?  Yes!  Get around to?  Not yet!


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Factor-talk mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/factor-talk

Reply via email to