Details
-
Bug
-
Resolution: Invalid
-
Medium
-
None
-
3.11
-
None
-
3.11
-
Description
The current code for MappedSortComparator has some unwanted hard memory references to an array of strings.
final String field = fieldname.intern();
final Object[] cachedValues = getLuceneValues(field, reader);
final Comparator comparator = sorter.getComparator();
return new ScoreDocComparator()
{
public int compare(ScoreDoc i, ScoreDoc j)
{
Object value1 = cachedValues[i.doc];
Object value2 = cachedValues[j.doc];
Since Lucene caches the results it means that the memory footprint of JIRA is much bigger than it could be. Especially over time as more searches are done.