Github user jpountz commented on a diff in the pull request:
https://github.com/apache/lucene-solr/pull/525#discussion_r243649022
--- Diff:
lucene/test-framework/src/java/org/apache/lucene/index/BaseDocValuesFormatTestCase.java
---
@@ -1204,6 +1204,12 @@ public void testRandomSortedBytes() throws
IOException {
}
private void doTestNumericsVsStoredFields(double density, LongSupplier
longs) throws Exception {
+ doTestNumericsVsStoredFields(density, longs, 256);
+ // TODO: 200000 docs are needed to test jumps properly (see
LUCENE-8585), but that is quite slow (few minutes).
+ // Maybe it can be nightly?
+ //doTestNumericsVsStoredFields(density, longs, 200000);
--- End diff --
By the way if we want to test such large numbers of documents, we should
avoid RandomIndexWriter (use IndexWriter instead) and set reasonable values of
maxBufferedDocs+RAMSizeBufferMB+mergePolicy instead of relying on random values
that are set by LuceneTestCase#newIndexWriterConfig and make indexing slow (but
are useful to increase coverage of flushing+merging)
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
