[ https://issues.apache.org/jira/browse/OAK-10384?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17763583#comment-17763583 ]
Amit Jain commented on OAK-10384: --------------------------------- bq. Is this something that should go into 1.22? No, the improvement introduced with - https://issues.apache.org/jira/browse/OAK-9707 hasn't been backported. > Fix stripping of large indexed ordered properties > ------------------------------------------------- > > Key: OAK-10384 > URL: https://issues.apache.org/jira/browse/OAK-10384 > Project: Jackrabbit Oak > Issue Type: Bug > Components: lucene > Reporter: Amit Jain > Assignee: Amit Jain > Priority: Major > Fix For: 1.58.0 > > > Currently, the ordered indexed properties are truncated at the max length > supported by lucene at 32766 in the > [LuceneDocumentMaker|https://github.com/apache/jackrabbit-oak/blob/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LuceneDocumentMaker.java#L290-L294]. > The problem is lucene uses a class {{BytesRef}} to represent strings which > converts it to UTF-8. It then uses the length from this converted string to > enforce the limit. The transformation between java unicode string to utf-8 > can cause the length to increase for non-ascii characters. -- This message was sent by Atlassian Jira (v8.20.10#820010)