hadoop-mapreduce-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Colin P. McCabe" <cmcc...@apache.org>
Subject Re: 2.7 status
Date Thu, 19 Feb 2015 07:09:27 GMT
On Wed, Feb 18, 2015 at 1:49 AM, Steve Loughran <stevel@hortonworks.com> wrote:
> On 17 February 2015 at 21:52:47, Colin McCabe (cmccabe@alumni.cmu.edu<mailto:cmccabe@alumni.cmu.edu>)
> Re: building Windows binaries. Do we release binaries for all the
> Linux and UNIX architectures? I thought we didn't. It seems a little
> inconsistent to release binaries just for Windows, but not for those
> other architectures and OSes. I wonder if we can improve this
> situation?
> Windows has less variance; one CPU family; if you target one version its there; if you
test on that then you can say "works on 64-bit server 2012". Win32 isn't handled no-one would
ever use it in production (RAM limitations alone). There's clearly interest in JIRA for a
version of the client-side code there
> There's a more subtle detail: we don't expect all windows users to have the build tools.
Whereas on Linux it's pretty much all there.
> Now, if someone wanted to produce native libs for other platforms, and help test them,
we'd probably have to think about what to do then.

I guess one more issue is that Windows requires the native bits,
whereas Linux does not.  I suppose that's a good argument for doing

> On that topic, Colin -can you take a look at  https://issues.apache.org/jira/browse/HADOOP-10846
; PowerPC native CRC. I Think the latest patch isn't going to have any adverse effects on
x86, but more native-code reviews would help.

I'll try to take a look tomorrow or maybe Friday.  Long backlog right now.


View raw message