[ 
https://issues.apache.org/jira/browse/SOLR-10926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16101923#comment-16101923
 ] 

Steve Rowe commented on SOLR-10926:
-----------------------------------

If there are no objections, tomorrow I'll commit the patch below to increase 
randomized points/trie testing ratio from 1:1 to 4:1. I plan on making this 
ratio the same on master, branch_7x and branch_7_0.

{noformat}
diff --git a/solr/test-framework/src/java/org/apache/solr/SolrTestCaseJ4.java 
b/solr/test-framework/src/java/org/apache/solr/SolrTestCaseJ4.java
index 6e23d45d30..0446093324 100644
--- a/solr/test-framework/src/java/org/apache/solr/SolrTestCaseJ4.java
+++ b/solr/test-framework/src/java/org/apache/solr/SolrTestCaseJ4.java
@@ -2705,7 +2705,7 @@ public abstract class SolrTestCaseJ4 extends 
LuceneTestCase {
     System.setProperty(NUMERIC_DOCVALUES_SYSPROP, ""+useDV);
     
     // consume a consistent amount of random data even if sysprop/annotation 
is set
-    final boolean randUsePoints = random().nextBoolean();
+    final boolean randUsePoints = 0 != random().nextInt(5);  // 80% likelihood
 
     final String usePointsStr = System.getProperty(USE_NUMERIC_POINTS_SYSPROP);
     final boolean usePoints = (null == usePointsStr) ? randUsePoints : 
Boolean.parseBoolean(usePointsStr);
{noformat}

> increase the odds of randomly choosing point fields in our SolrTestCaseJ4 
> numeric type randomization
> ----------------------------------------------------------------------------------------------------
>
>                 Key: SOLR-10926
>                 URL: https://issues.apache.org/jira/browse/SOLR-10926
>             Project: Solr
>          Issue Type: Test
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Hoss Man
>            Priority: Blocker
>              Labels: numeric-tries-to-points
>             Fix For: 7.0
>
>
> currently it's a 50/50 chance of using point fields vs trie fields ... once 
> we are more confident in the utility/reliability of point fields and/or they 
> are the "default" in our example configsets, we should tweak those odds so 
> Point fields get tested much more often then TrieFields



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to