I guess it is due to the difference in sort algorithm(?). Anyway,
sorting in indexSerial has not to be done because it can (and it was)
changes the order of the fields added to the document.
Andrei.
Digy (JIRA) wrote:
[ https://issues.apache.org/jira/browse/LUCENENET-260?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12779047#action_12779047 ]
Digy commented on LUCENENET-260:
--------------------------------
The problem with this issue is that although Lucene.Java uses also "Collections.sort(fields,
fieldNameComparator)" in "indexSerial"
this test fails only in Lucene.NET. What makes this difference?
DIGY
src\Test\Index\TestStressIndexing2.cs
-------------------------------------
Key: LUCENENET-260
URL: https://issues.apache.org/jira/browse/LUCENENET-260
Project: Lucene.Net
Issue Type: Bug
Reporter: Iliev Andrei
Assignee: Digy
Attachments: TestStressIndexing2.patch
Dead lock issues and LUCENENET-143.
Upd: Dead lock has been resolve in LUCENENET-255. So I update the patch.
