[
https://issues.apache.org/jira/browse/LUCENE-4873?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stefan Pohl updated LUCENE-4873:
--------------------------------
Attachment: LUCENE-4873.patch
I screwed up the implementation of the heap operation that I added
(minheapRemove), so actually no extreme corner case, it only didn't pop up
because this method is called rarely.
Patch attached (sorry, I also reformatted and cleaned up up some comments in
the same patch). I added a method to check the validity of the heap
(minheapCheck) that I call within asserts after every relevant operation,
please remove these expensive operation calls if you think that assertions
should be fast because many users run with assertions enabled and could be
irritated.
I couldn't produce any other test failings, and lucenebench shows the same
results.
This just exemplifies that one shouldn't re-implement basic data structures in
each and every class.
Would it make sense to add heap operations to e.g. ArrayUtil and refactor the
codebase? Or is it known that this would mean prohibitive performance impact?
> corner case in MinShouldMatchSumScorer when there are many terms
> ----------------------------------------------------------------
>
> Key: LUCENE-4873
> URL: https://issues.apache.org/jira/browse/LUCENE-4873
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/query/scoring
> Affects Versions: 4.3
> Reporter: Robert Muir
> Attachments: LUCENE-4873.patch
>
>
> I think this bug is some extreme corner case...
> This test currently only uses up to 9 terms. By increasing it to 26 and
> blasting the test, I was able to uncover a bug.
> Here's the seed: ant test -Dtestcase=TestMinShouldMatch2
> -Dtests.method=testNextAllTerms -Dtests.seed=E0334C37E6E190D8
> -Dtests.slow=true -Dtests.locale=pl_PL -Dtests.timezone=Asia/Thimphu
> -Dtests.file.encoding=US-ASCII
> Here's the patch to make the test use 26 terms.
> {noformat}
> Index: lucene/core/src/test/org/apache/lucene/search/TestMinShouldMatch2.java
> ===================================================================
> --- lucene/core/src/test/org/apache/lucene/search/TestMinShouldMatch2.java
> (revision 1459937)
> +++ lucene/core/src/test/org/apache/lucene/search/TestMinShouldMatch2.java
> (working copy)
> @@ -56,7 +56,7 @@
> static final String alwaysTerms[] = { "a" };
> static final String commonTerms[] = { "b", "c", "d" };
> static final String mediumTerms[] = { "e", "f", "g" };
> - static final String rareTerms[] = { "h", "i", "j" };
> + static final String rareTerms[] = { "h", "i", "j", "k", "l", "m", "n",
> "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z" };
>
> @Override
> public void setUp() throws Exception {
> {noformat}
--
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: [email protected]
For additional commands, e-mail: [email protected]
