[ https://issues.apache.org/jira/browse/UIMA-6199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17055071#comment-17055071 ]
Richard Eckart de Castilho commented on UIMA-6199: -------------------------------------------------- DKPro Core does not use the default UIMA DocumentAnnotation. It has it's own DocumentMetaData annotation which inherits from DocumentAnnotation. If the DKPro Core XmiReader finds a UIMA DocumentAnnotation, it removes it and adds instead a DocumentMetaData annotation. > Unindexed document annotation added back to index during deserialization > ------------------------------------------------------------------------ > > Key: UIMA-6199 > URL: https://issues.apache.org/jira/browse/UIMA-6199 > Project: UIMA > Issue Type: Bug > Components: Core Java Framework > Affects Versions: 3.1.1SDK > Reporter: Richard Eckart de Castilho > Priority: Major > Fix For: 3.2.0SDK > > Time Spent: 10m > Remaining Estimate: 0h > > If a CAS contains a document annotation that was removed from the indexes, it > may be added back to the indexes during deserialization if > {{CASImpl.ALWAYS_HOLD_ONTO_FSS}} is {{true}}. > There is also a second scenario triggering this issue that works without > enabling {{CASImpl.ALWAYS_HOLD_ONTO_FSS}}. It involves creating an FS which > references the document annotation before removing it from the indexes. -- This message was sent by Atlassian Jira (v8.3.4#803005)