Resolve findbugs warnings in core/streaming/pipes/examples

• Key: HADOOP-5679
• Type: Bug
• Status: Resolved
• Resolution: Fixed
• Priority: Major
• Assignee: Jothi Padmanabhan
• Reporter: Jothi Padmanabhan
• Fix Version/s: 0.21.0

Description

Towards a solution for HADOOP-5628, we need to resolve all findbugs warnings. This jira will try to resolve the findbugs warnings where ever possible and suppress them where resolution is not possible.

Comments

[Jothi Padmanabhan]

Patch for review

[Sharad Agarwal]

Patch looks good.
Minor nit: The signature of public method ProgramDriver#driver(String[] args) has changed. So should we make this as incompatible change ?

[Jothi Padmanabhan]

Ant tests passed on my local box.

Test Patch result

[exec] -1 overall.
[exec]
[exec] +1 @author. The patch does not contain any @author tags.
[exec]
[exec] +1 tests included. The patch appears to include 3 new or modified tests.
[exec]
[exec] +1 javadoc. The javadoc tool did not generate any warning messages.
[exec]
[exec] +1 javac. The applied patch does not increase the total number of javac compiler warnings.
[exec]
[exec] +1 findbugs. The patch does not introduce any new Findbugs warnings.
[exec]
[exec] +1 Eclipse classpath. The patch retains Eclipse classpath integrity.
[exec]
[exec] -1 release audit. The applied patch generated 481 release audit warnings (more than the trunk's current 474 warnings).

[Sharad Agarwal]

I just committed this. Thanks Jothi.

[Hudson]

Integrated in Hadoop-trunk #830 (See http://hudson.zones.apache.org/hudson/job/Hadoop-trunk/830/)
. Resolve findbugs warnings in core/streaming/pipes/examples. Contributed by Jothi Padmanabhan.

[Jordà Polo]

Running local jobs seems to fail for me since this patch was applied ("Mkdirs failed to create...").

The new code checks if either mkdirs() on hadoop.tmp.dir failed OR if it isn't a directory... but it doesn't take into account that the directory may already exist. If the directory existed before, it will always fail and exit since mkdirs() only returns "true" if it is able to create the directory.

Wouldn't an AND be more appropriate here? I mean (!b && !tmpDir.isDirectory()) instead of (!b || !tmpDir.isDirectory()) in src/core/org/apache/hadoop/util/RunJar.java, line 111.

EDIT: The original code (before applying hadoop-5679.patch) looks fine to me too: (!tmpDir.isDirectory()).

[Jothi Padmanabhan]

Yes, this is a bug. HADOOP-5809 resolves this.

[Robert Chansler]

Editorial pass over all release notes prior to publication of 0.21. This is just routine.