lucenenet-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christopher Currens <currens.ch...@gmail.com>
Subject Re: Spatial tests failure
Date Sun, 23 Sep 2012 03:52:55 GMT
The problem is line 43 of CompatibilityExtensions.cs.  Whenever you
have a problem that only occurs when multiple tests are run, it's
generally safe to say you have a problem with a static member.

The problem is that twodoubles__x and twodoubles__y are created during
the twodoubles strategies on any of the other test suites, but they
set it to a FixedBitSet, where testIntersections uses a
MatchAllBitSet...the index is out of range because the other tests add
three documents, where testIntersections adds 13.  Since the others
are cached, it gets to the fourth document and errors out.

I'm not terribly familiar with the spatial libraries, but I can't find
this class in the java code.  Is this something we've added for
compatibility with the Java code?


On Sat, Sep 22, 2012 at 7:41 PM, Prescott Nasser <geobmx540@hotmail.com> wrote:
> Ill also take a look in a couple of hours.
>
> Sent from my Windows Phone
> ________________________________
> From: Christopher Currens
> Sent: 9/22/2012 5:35 PM
> To: dev@lucenenet.apache.org
> Subject: Re: Spatial tests failure
>
> I can look at this in a couple hours, I'm not near a computer right now.
> On Sep 22, 2012 4:37 PM, "Itamar Syn-Hershko" <itamar@code972.com> wrote:
>
>> Guys, I need an extra set of eyes on the following.
>>
>> In Contrib.Spatial.Test all test pass, except when they are being run as a
>> suite (I do this with TestDriven.NET, clicking "Run tests" on the tests
>> project), then there is one test that is failing:
>>
>> testIntersections in PortedSolr3Test.cs , with the "twodoubles" strategy.
>>
>> It all passes consistently when running it directly. It is only when I'm
>> running it as a suite where it fails.
>>
>> In the suite it runs after the tests in TestBBoxStrategy. I made sure
>> everything is being cleaned up properly, and still it happens.
>>
>> The failure itself is due to a non-existent doc-id being looked up by the
>> core.
>>
>> I'm probably missing something very stupid!
>>

Mime
View raw message