commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thomas Neidhart <thomas.neidh...@gmail.com>
Subject Re: [continuum] BUILD FAILURE: Apache Commons - Commons Math - Default Maven 2 Build Definition (Java 1.5)
Date Fri, 03 Feb 2012 12:06:38 GMT
On 02/02/2012 11:57 AM, Gilles Sadowski wrote:
> On Thu, Feb 02, 2012 at 12:40:14AM +0000, Continuum@vmbuild wrote:
> 
> The failure is unrelated to the commit, but has to do with this issue:
>   https://issues.apache.org/jira/browse/MATH-588
> 
> I don't know whether the test tolerance is too stringent or there really is
> a bug. Could someone have a look?

Hi,

I have looked into this issue, and imo the test tolerance is too stringent:

>> MeanTest
>>     testWeightedConsistency :
>>   java.lang.AssertionError
>>   java.lang.AssertionError: expected:<-5.261475053036868E-4> but was:<-5.26147505303894E-4>

The tests compares the two computed means with a relative error of
10-14, which can lead to an quite small absolute error in case the
randomized test values are distributed in an unfortunate way.

In the failing example, the mean is calculated to be -5.261e-4 which
leads to an absolute error of 10-18 while the two means are equal up to
10-15.

The bug report is very similar:

the expected mean is ~2.2e-3 and the values are equal up to 10-15.

Thomas

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org


Mime
View raw message