Since the build.sh script is failing on one platform due to known reasons which can be reproduced (assuming Mike's fix works for Bill), I don't see a reason to push back the 1.4.2 release.

Given that this hasn't been addressed in RC3, what does everyone else think? Is providing a notice along with the 1.4.2 release sufficient to warn people about the known rat issue on OSX and then create a fix for 1.4.3 or 1.5.0 sufficient? Do we even feel the need to note the workaround (as this would require another round of signing, staging and voting?

Personally, I don't think we need to revisit this for 1.4.2 as long as Mike's fix works for Bill because most users aren't going to use the build.sh script anyways. They'll download the source just run `mvn package && mvn assembly:single -N` as the README tells them, or they'll just download the pre-build distribution.

On 10/22/2012 10:53 PM, William Slacum wrote:
Thanks Mr. Moustache-- I ended up just kicking up the expected number of
files from 53 to 55. Should we put in an OS check in the script as a band
aid, since it seems OSX isn't playing nicely?

On Mon, Oct 22, 2012 at 10:41 PM, Michael Flester <[email protected]> wrote:

Wilhelm --

For the release of 1.3.6 a while back if I switched out the "mvn
rat:check" line in
assemble/build.sh with this:

   mvn org.apache.rat:apache-rat-plugin:check

I could get the build to pass on OS X, otherwise the odp
files would break it like you say. I just tested this on
trunk and it did not make the build pass but I don't think
the issue on trunk is related to odp files.


On Mon, Oct 22, 2012 at 10:34 PM, William Slacum <
[email protected]> wrote:

I replied to the thread David made, since I think Billie has run into the
same issue. I'm on OSX 10.7.5 and I believe
it's docs/src/developer_manual/component_docs.odp
and test/system/continuous/ScaleTest.odp, since she mentioned that some
versions seem not care if it's binary or not.


On Mon, Oct 22, 2012 at 10:09 PM, Eric Newton <[email protected]>wrote:

Can you identify a file that is missing a license or has an incorrect
license?

I have run the build on RHEL 6, and Ubuntu 12.04.  In what environment
does the build fail?

-Eric


On Mon, Oct 22, 2012 at 10:06 PM, William Slacum <
[email protected]> wrote:

-1, since I'm running into the rat issue reported by Dave Medinets when
running build.sh.


On Mon, Oct 22, 2012 at 12:20 PM, Keith Turner <[email protected]>wrote:

On Mon, Oct 22, 2012 at 9:52 AM, Josh Elser <[email protected]>
wrote:
I agree. If it's not a quick fix, we should just revert the change
and fix
it properly in the next release.

Since this is a bug introduced in 1.4.1, Christopher suggested rolling
back the changes made in 1.4.1 in the ticket.  I like this idea and
will take a stab at it today.



On 10/19/12 5:32 PM, Christopher Tubbs wrote:

I don't know that ACCUMULO-826 should be fixed before release, as
I'm
not sure there's a good fix without changing the API, and these
issues
may occur in several places in the MapReduce API.

--
Christopher L Tubbs II
http://gravatar.com/ctubbsii


On Fri, Oct 19, 2012 at 3:41 PM, Eric Newton <[email protected]

wrote:

I agree.  And thanks for taking some time to test the candidate.

It would be great if we could get some feedback from all the
committers,
and
soon. I assume many of them will be busy in NY next week.

If you look at the CHANGES for 1.4.2, there are some significant
bug
fixes.
We want to make sure the final release doesn't contain any
unexpected
surprises like this.

-Eric

On Fri, Oct 19, 2012 at 3:33 PM, Keith Turner <[email protected]>
wrote:

While testing 1.4.2rc2, I ran into ACCUMULO-826.   I think this
is a
pretty severe issue that occurs under a fairly common use case.
It
sucks to have your M/R job die after a few hours because you
killed
the processes that started the job.   I am thinking its worth
holding
1.4.2 up in order to fix this issue, thoughts?

On Thu, Oct 18, 2012 at 12:46 PM, Eric Newton <
[email protected]>
wrote:

Please vote on releasing the following candidate as Apache
Accumulo
version 1.4.2.

The src tar ball was generated by exporting:

https://svn.apache.org/repos/asf/accumulo/tags/1.4.2rc2

To build the dist tar ball from the source run the following
command:
     src/assemble/build.sh

Tarballs, checksums, signatures:
    http://people.apache.org/~ecn/1.4.2rc2

Maven Staged Repository:



https://repository.apache.org/content/repositories/orgapacheaccumulo-135

Keys:
    http://www.apache.org/dist/accumulo/KEYS

Changes:

https://svn.apache.org/repos/asf/accumulo/tags/1.4.2rc2/CHANGES

The vote will be held open for the next 72 hours.

The only change from RC1 was ACCUMULO-823.










Reply via email to