[
https://issues.apache.org/jira/browse/HBASE-18838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16291710#comment-16291710
]
Appy commented on HBASE-18838:
------------------------------
yeah.
+1 on v4.
The question of whether parent exclusions are overwritten or added remains
mystery. I guess you don't have definite word on it either, except what you're
observing.
While uneasy, it doesn't warrant blocking patch.
Since it's deps change, feel free to commit
- since you've verified tarball
- and after testing shaded-checks locally with hadoop3 & successful hadoopcheck
from QA (will suggest adding single line change to hbase-common if you want to
bring down pre-commit runtime from 4hr to 45 min)
> shaded artifacts are incorrect when built against hadoop 3
> ----------------------------------------------------------
>
> Key: HBASE-18838
> URL: https://issues.apache.org/jira/browse/HBASE-18838
> Project: HBase
> Issue Type: Bug
> Components: Client
> Affects Versions: 2.0.0-alpha-3
> Reporter: Sean Busbey
> Assignee: Mike Drob
> Priority: Critical
> Fix For: 2.0.0-beta-1
>
> Attachments: HBASE-18838-WIP.v2.patch, HBASE-18838.WIP.patch,
> HBASE-18838.v3.patch, HBASE-18838.v4.patch
>
>
> Building master/branch-2 against the hadoop-3 profile results in
> check-invariants screaming about unrelocated dependencies. will list details
> in comment.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
