[
https://issues.apache.org/jira/browse/HADOOP-12714?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15102038#comment-15102038
]
Colin Patrick McCabe commented on HADOOP-12714:
-----------------------------------------------
Ah, good catch. I assumed this was a file generated by glibc (since it's named
"testGlibCBugSpill"), but actually it's a (very large) resource checked into
the source tree. It looks like it gets copied from
{{./hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-nativetask/src/main/native/testData/testGlibCBugSpill}}
to
{{./hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-nativetask/target/native/test/testData/testGlibCBugSpill}}
as part of the build process. It wasn't being found because the cmakebuilder
plugin was running the test with the wrong working directory. I added a fix
for this in v2.
> Fix hadoop-mapreduce-client-nativetask unit test which fails when glibc is
> not buggy
> ------------------------------------------------------------------------------------
>
> Key: HADOOP-12714
> URL: https://issues.apache.org/jira/browse/HADOOP-12714
> Project: Hadoop Common
> Issue Type: Bug
> Components: native
> Affects Versions: 3.0.0
> Reporter: Colin Patrick McCabe
> Assignee: Colin Patrick McCabe
> Attachments: HADOOP-12714.001.patch
>
>
> Fix hadoop-mapreduce-client-nativetask unit test which fails when glibc is
> not buggy. It attempts to open a "glibc bug spill" file which doesn't exist
> unless glibc has a bug.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
