[
https://issues.apache.org/jira/browse/LUCENE-5222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13770831#comment-13770831
]
Robert Muir commented on LUCENE-5222:
-------------------------------------
{quote}
If I add the last two boolean values explicitly as true (doDocScores and
doMaxScore) to both search() and searchAfter() calls, the test succeeds
{quote}
Indeed: I just tested this and it works (see patch below)
Additionally i noticed TestSearchAfter in core somehow "hides the bug" with
this stuff in its test:
{noformat}
if (sort == Sort.RELEVANCE) {
paged = searcher.searchAfter(lastBottom, query, filter, pageSize,
sort, true, doMaxScore);
{noformat}
So there is a bug here in searchAfter/TopFieldCollector, and i'm also unhappy
TestSearchAfter is not testing multiple sorts!
{noformat}
Index: src/test/org/apache/lucene/expressions/TestExpressionSorts.java
===================================================================
--- src/test/org/apache/lucene/expressions/TestExpressionSorts.java
(revision 1524082)
+++ src/test/org/apache/lucene/expressions/TestExpressionSorts.java
(working copy)
@@ -131,7 +131,10 @@
void assertQuery(Query query, Filter filter, Sort sort) throws Exception {
int size = _TestUtil.nextInt(random(), 1,
searcher.getIndexReader().maxDoc()/5);
- TopDocs expected = searcher.search(query, filter, size, sort,
random().nextBoolean(), random().nextBoolean());
+ // consume randomness
+ random().nextBoolean();
+ random().nextBoolean();
+ TopDocs expected = searcher.search(query, filter, size, sort, true, false);
// make our actual sort, mutating original by replacing some of the
// sortfields with equivalent expressions
@@ -152,12 +155,15 @@
}
Sort mutatedSort = new Sort(mutated);
- TopDocs actual = searcher.search(query, filter, size, mutatedSort,
random().nextBoolean(), random().nextBoolean());
+ // consume randomness
+ random().nextBoolean();
+ random().nextBoolean();
+ TopDocs actual = searcher.search(query, filter, size, mutatedSort, true,
false);
CheckHits.checkEqual(query, expected.scoreDocs, actual.scoreDocs);
if (size < actual.totalHits) {
- expected = searcher.searchAfter(expected.scoreDocs[size-1], query,
filter, size, sort);
- actual = searcher.searchAfter(actual.scoreDocs[size-1], query, filter,
size, mutatedSort);
+ expected = searcher.searchAfter(expected.scoreDocs[size-1], query,
filter, size, sort, true, false);
+ actual = searcher.searchAfter(actual.scoreDocs[size-1], query, filter,
size, mutatedSort, true, false);
CheckHits.checkEqual(query, expected.scoreDocs, actual.scoreDocs);
}
}
{noformat}
> TestExpressionSorts fails sometimes when using expression returning score
> -------------------------------------------------------------------------
>
> Key: LUCENE-5222
> URL: https://issues.apache.org/jira/browse/LUCENE-5222
> Project: Lucene - Core
> Issue Type: Bug
> Reporter: Ryan Ernst
>
> Jenkins picked this up. Repeat with:
> {code}
> ant test -Dtestcase=TestExpressionSorts -Dtests.method=testQueries
> -Dtests.seed=115AD00ED89D9F7B -Dtests.multiplier=3 -Dtests.slow=true
> -Dtests.locale=no_NO -Dtests.timezone=America/Nassau
> -Dtests.file.encoding=US-ASCII
> {code}
> It appears to have to do with scoring, as removing the score sort from the
> original sorts causes the tests to pass. If you remove the possible
> discrepancy between doDocScores and docMaxScore params to searcher.search,
> then the test gets farther before failing.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
