Hey,
Looks like there are some memory leaks in JArray_string, when
iterating or accessing its items. I've seen it reproduce with 2.4.1
on linux and osx. Attached is a test case to demonstrate it using
Document.getValues, and a potential patch.
"""
Memory leak in string array, as returned by Document.getValues.
Low maxheap set to introduce error quickly.
"""
import itertools
import lucene
if __name__ == '__main__':
lucene.initVM(lucene.CLASSPATH, maxheap='10m')
writer = lucene.IndexWriter(lucene.RAMDirectory(), lucene.StandardAnalyzer())
doc = lucene.Document()
for value in ('one', 'two'):
doc.add(lucene.Field('test', value, lucene.Field.Store.YES, lucene.Field.Index.NO))
writer.addDocument(doc)
writer.commit()
reader = lucene.IndexReader.open(writer.directory)
for i in itertools.count():
values = reader.document(0).getValues('test')
# comment out loop through values, and the leak disappears
for value in values:
pass
if not i % 10000:
print i
