[
https://issues.apache.org/jira/browse/HADOOP-4280?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12634695#action_12634695
]
Pete Wyckoff commented on HADOOP-4280:
--------------------------------------
Sorry - now see that part of the code is where connected as super user and the
file is 0666 so no problem there.
It is weird that the last modification time of the file is coming back as
negative. The time in FileStatus is a long in ms and hdfs.c converts that to a
tTime by diving it to 1000. The result in JNI is treated as a long (J type) but
then there''s a cast to a tTime:
{code}
if (invokeMethod(env, &jVal, &jExc, INSTANCE, jStat,
HADOOP_STAT, "getModificationTime", "()J") != 0) {
errno = errnoFromException(jExc, env, "org.apache.hadoop.fs."
"FileStatus::getModificationTime");
return -1;
}
fileInfo->mLastMod = (tTime) (jVal.j) / 1000;
{code}
Which is weird since it should be:
{code}
fileInfo->mLastMod = (tTime) (jVal.j / 1000);
{code}
That may be causing the problem. I am using gcc 4.0.1 and not having a problem,
but ...
> test-libhdfs consistently fails on trunk
> ----------------------------------------
>
> Key: HADOOP-4280
> URL: https://issues.apache.org/jira/browse/HADOOP-4280
> Project: Hadoop Core
> Issue Type: Bug
> Components: libhdfs
> Affects Versions: 0.19.0
> Reporter: Raghu Angadi
> Assignee: Pete Wyckoff
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: test-libhdfs-failure.txt
>
>
> {{ant -Dlibhdfs=yes test-libhdfs}} fails on 0.19 and trunk. It is ok on 0.18.
> I will attach output from a run.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
