Zongheng pointed out in my SPARK-3329 PR (https://github.com/apache/spark/pull/2220) that Aaron had already fixed this issue but that it had gotten inadvertently clobbered by another patch. I don't know how the project handles this kind of problem, but I've rewritten my SPARK-3329 branch to cherry-pick Aaron's fix (also fixing a merge conflict and handling a test case that it didn't).
The other weird spurious testsuite failures related to orderings I've seen were in "DESCRIBE FUNCTION EXTENDED" for functions with lists of synonyms (e.g. STDDEV). I can't reproduce those now but will take another look later this week. best, wb ----- Original Message ----- > From: "Sean Owen" <so...@cloudera.com> > To: "Will Benton" <wi...@redhat.com> > Cc: "Patrick Wendell" <pwend...@gmail.com>, dev@spark.apache.org > Sent: Sunday, August 31, 2014 12:18:42 PM > Subject: Re: [VOTE] Release Apache Spark 1.1.0 (RC3) > > Fantastic. As it happens, I just fixed up Mahout's tests for Java 8 > and observed a lot of the same type of failure. > > I'm about to submit PRs for the two issues I identified. AFAICT these > 3 then cover the failures I mentioned: > > https://issues.apache.org/jira/browse/SPARK-3329 > https://issues.apache.org/jira/browse/SPARK-3330 > https://issues.apache.org/jira/browse/SPARK-3331 > > I'd argue that none necessarily block a release, since they just > represent a problem with test-only code in Java 8, with the test-only > context of Jenkins and multiple profiles, and with a trivial > configuration in a style check for Python. Should be fixed but none > indicate a bug in the release. > > On Sun, Aug 31, 2014 at 6:11 PM, Will Benton <wi...@redhat.com> wrote: > > ----- Original Message ----- > > > >> dev/run-tests fails two tests (1 Hive, 1 Kafka Streaming) for me > >> locally on 1.1.0-rc3. Does anyone else see that? It may be my env. > >> Although I still see the Hive failure on Debian too: > >> > >> [info] - SET commands semantics for a HiveContext *** FAILED *** > >> [info] Expected Array("spark.sql.key.usedfortestonly=test.val.0", > >> "spark.sql.key.usedfortestonlyspark.sql.key.usedfortestonly=test.val.0test.val.0"), > >> but got > >> Array("spark.sql.key.usedfortestonlyspark.sql.key.usedfortestonly=test.val.0test.val.0", > >> "spark.sql.key.usedfortestonly=test.val.0") (HiveQuerySuite.scala:541) > > > > I've seen this error before. (In particular, I've seen it on my OS X > > machine using Oracle JDK 8 but not on Fedora using OpenJDK.) I've also > > seen similar errors in topic branches (but not on master) that seem to > > indicate that tests depend on sets of pairs arriving from Hive in a > > particular order; it seems that this isn't a safe assumption. > > > > I just submitted a (trivial) PR to fix this spurious failure: > > https://github.com/apache/spark/pull/2220 > > > > > > best, > > wb > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org > For additional commands, e-mail: dev-h...@spark.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org For additional commands, e-mail: dev-h...@spark.apache.org