[
https://issues.apache.org/jira/browse/CALCITE-6895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17936034#comment-17936034
]
Julian Hyde commented on CALCITE-6895:
--------------------------------------
It was I who put in the previous test. I wanted to sanity check - make sure
that SOME version is coming through - without adding to the burden of every
release. I still think it was the right compromise.
We single-source the versions, so if some number is coming through, it is very
likely the right one. But we periodically change how we roll jar files, and the
chances of the version number getting lost during those changes is high enough
that we need to test for it.
> Change JdbcTest.testVersion to strictly match minor version in
> gradle.properties
> --------------------------------------------------------------------------------
>
> Key: CALCITE-6895
> URL: https://issues.apache.org/jira/browse/CALCITE-6895
> Project: Calcite
> Issue Type: Task
> Reporter: Stamatis Zampetakis
> Assignee: Stamatis Zampetakis
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.40.0
>
>
> A [recent
> bump|https://github.com/apache/calcite/commit/f0195df7fe18b8d99d5e1d259a3e35834e4c77e2]
> of calcite.version to 1.40.0 broke {{JdbcTest.test}} version cause the
> following condition was no longer true.
> {code:java}
> assertTrue(driverMinor >= 0 && driverMinor < 40);
> {code}
> If we want to test that versions in properties really match those in JDBC it
> makes more sense to have strict equality instead of range. The drawback is
> that we have to update the test after every version bump.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
