On Nov 25, 2009, at 1:46, "Uwe Schindler" <[email protected]> wrote:
I would really not see this as a real test failure. The test assumes
that a
compound file index is slower during indexing, which is normally so.
But
maybe the scheduler of your O/S did something strange at the time
the test
ran and made it take longer to index without compound file. Is it
reproducible?
In my opinion, the test is invalid, it just shows something that
happens
most of the time, but asserting it is wrong.
I looked at the Java Version of the test:
http://java.codefetch.com/example/in/LuceneInAction/src/lia/indexing/Compoun
dVersusMultiFileIndexTest.java
Yes, that is correct. The assertion is a little optimistic. Its
failure can be ignored.
Andi..
Uwe
-----
Uwe Schindler
H.-H.-Meier-Allee 63, D-28213 Bremen
http://www.thetaphi.de
eMail: [email protected]
-----Original Message-----
From: Helmut Jarausch [mailto:[email protected]]
Sent: Wednesday, November 25, 2009 10:33 AM
To: [email protected]
Cc: Andi Vajda
Subject: Re: [VOTE] Release Apache Lucene Java 3.0.0 (take #2)
On 24 Nov, Andi Vajda wrote:
I built PyLucene trunk using Lucene Java source from the
artifact's svn
rev
and all unit tests and ported "Lucene in Action" tests pass.
+1 !
Andi..
Many thanks, Andi.
Here (AMD64, python-2.6.4) a single test fails
java development with ant: 0.502968132496
junit in action: 0.812919199467
0.812919199467: JUnit in Action
0.502968132496: Java Development with Ant
/usr/bin/python samples/LuceneInAction/
CompoundVersusMultiFileIndexTest.py
F
===
===================================================================
FAIL: testTiming
(lia.indexing.CompoundVersusMultiFileIndexTest.CompoundVersusMultiFileInde
xTest)
---
-------------------------------------------------------------------
Traceback (most recent call last):
File "/Work1/Obj/Python/pylucene-
build/samples/LuceneInAction/lia/indexing/
CompoundVersusMultiFileIndexTest
.py", line 62, in testTiming
self.assert_(cTiming > mTiming)
AssertionError
Hopefully, this isn't critical.
What could be the reason?
Thanks again,
Helmut.
--
Helmut Jarausch
Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany
