Github user acogoluegnes commented on a diff in the pull request:
https://github.com/apache/maven-surefire/pull/171#discussion_r153745052
--- Diff:
maven-surefire-common/src/test/java/org/apache/maven/plugin/surefire/SurefirePropertiesTest.java
---
@@ -70,9 +70,16 @@ public void testConstructWithOther()
src.setProperty( "b", "2" );
SurefireProperties orderedProperties = new SurefireProperties( src
);
// Cannot make assumptions about insertion order
- assertEquals( 2, orderedProperties.size() );
-
-
+ int expectedCount = 0;
+ // keys() uses the items property, more reliable to test than
size(),
+ // which is based on the Properties class
+ // see https://issues.apache.org/jira/browse/SUREFIRE-1445
+ Enumeration<Object> keys = orderedProperties.keys();
+ while (keys.hasMoreElements()) {
+ keys.nextElement();
+ expectedCount++;
+ }
+ assertEquals( 2, expectedCount );
--- End diff --
I don't get it. I used `keys()` on purpose because it relies on `items`,
which was not in-sync in the first place. `getStringKeySet()` relies
`keySet()`, which uses the base class. Both would work, but `getStringKeySet()`
doesn't catch the bug. Does that make sense?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
