hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan Burlison (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-11505) Various native parts use bswap incorrectly and unportably
Date Tue, 19 Jan 2016 22:23:39 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-11505?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15107572#comment-15107572

Alan Burlison commented on HADOOP-11505:

Yes, I think splitting the work up is a good idea.

I've done a little more digging and to do anything sophisticated with ant inside a pom.xml
is not possible, for example only running cmake if CMakeFiles.txt has been modified. That's
because you can only have one ant <target> block within a maven-antrun-plugin section
in the pom.xml. The recommended approach is to to move any complicated ant logic into a separate
build.xml file and just call ant from maven. I think such a refactoring would also make it
easier to develop/test the native build parts in isolation from the rest of the maven build.
The one wrinkle in this case is the use of the native-maven-plugin in the pom.xml to run javah
on the classfiles and generate the corresponding C headers. I've completely failed to find
any documentation for it so I've no idea if it is smart enough to only regenerate the .h when
the corresponding classfile is updated. However if the decision is to refactor the JNI build
into an ant build.xml then I think the javah parts should be moved into ant as well.

> Various native parts use bswap incorrectly and unportably
> ---------------------------------------------------------
>                 Key: HADOOP-11505
>                 URL: https://issues.apache.org/jira/browse/HADOOP-11505
>             Project: Hadoop Common
>          Issue Type: Bug
>    Affects Versions: 3.0.0
>            Reporter: Colin Patrick McCabe
>            Assignee: Alan Burlison
>             Fix For: 3.0.0
>         Attachments: HADOOP-11505.001.patch, HADOOP-11505.003.patch, HADOOP-11505.004.patch,
HADOOP-11505.005.patch, HADOOP-11505.006.patch, HADOOP-11505.007.patch
> hadoop-mapreduce-client-nativetask fails to use x86 optimizations in some cases.  Also,
on some alternate, non-x86, non-ARM architectures the generated code is incorrect.  Thanks
to Steve Loughran and Edward Nevill for finding this.

This message was sent by Atlassian JIRA

View raw message