[jira] [Commented] (CALCITE-2365) Calcite is not compiled after bumping Avatica version to 1.12.0

["Julian Hyde (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CALCITE-2365?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16514434#comment-16514434 ] 

Julian Hyde commented on CALCITE-2365:
--------------------------------------

It may or may not be related, but I'm seeing misc.iq and scalar.iq fail on Windows but not on macOS or Linux. It's not the usual problem with line endings, nor is it about JDK versions. For some reason some of the logic concerning decorrelation is working differently on Windows.

> Calcite is not compiled after bumping Avatica version to 1.12.0
> ---------------------------------------------------------------
>
>                 Key: CALCITE-2365
>                 URL: https://issues.apache.org/jira/browse/CALCITE-2365
>             Project: Calcite
>          Issue Type: Bug
>            Reporter: Sergey Nuyanzin
>            Assignee: Julian Hyde
>            Priority: Major
>
> I guess the main reason is CALCITE-2219.
> tried to fix it here https://github.com/apache/calcite/compare/master...snuyanzin:CALCITE_AVATICA_12_UPDATE
> however there are at least 2 issues which are not fixed but only workarounded
> # _org.apache.calcite.test.LatticeTest#testGroupByEmpty3_ connection pool factory is used however 3 methods (explainContains, and 2 returnsUnordered ) lead to _org.apache.calcite.test.CalciteAssert#assertQuery_ which closes connections without paying attention if its from pool or not
> # the same test. mats updated in explainplan and execute => that is why finally size is 6.
> # strange but only now the test from core/src/test/resources/sql/misc.iq starts failing. Logically because of difference in data it must have failed earlier
> it would be nice to have some advice at least relating to first 2 bullets to fix it in a better way 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)