> On May 9, 2014, 5:23 p.m., Sean Busbey wrote:
> > server/base/src/main/java/org/apache/accumulo/server/util/CustomNonBlockingServer.java,
> >  lines 36-38
> > <https://reviews.apache.org/r/21043/diff/1/?file=573991#file573991line36>
> >
> >     nit: whitespace
> 
> Christopher Tubbs wrote:
>     This is introduced by a known issue with the Eclipse formatter 
> (https://bugs.eclipse.org/bugs/show_bug.cgi?id=270745). I'm not too concerned.
> 
> Sean Busbey wrote:
>     Provided you can remove it manually, please do so.

You're asking me to diverge away from the implementation of an established 
formatting standard for the project. While I agree these trailing whitespace 
should not be there, I'd rather not spend time making manual modifications to 
formatting that we've agreed on as a standard. I prefer them gone as well, but 
it's not worth the time and effort to make manual modifications. If it's a 
blocker for you, I'll remove it (I assume it's not, since you described it as a 
"nit"), otherwise, I'd prefer to spend my time elsewhere.

I've accepted patches with these removed and with them present. I ask that you 
be tolerant to both as well, until there is a formatting standard that we've 
agreed on that fixes this issue once and for all, and focus the review effort 
on more important bits. Thanks.


> On May 9, 2014, 5:23 p.m., Sean Busbey wrote:
> > server/base/src/main/java/org/apache/accumulo/server/util/CustomNonBlockingServer.java,
> >  lines 38-41
> > <https://reviews.apache.org/r/21043/diff/1/?file=573991#file573991line38>
> >
> >     is there an upstream ticket in Thrift that would allow us to skip 
> > having our own implementation?
> 
> Christopher Tubbs wrote:
>     Yes, see the THRIFT issue linked in the JIRA, and the additional issues 
> linked from there. This issue is a regression that has occurred at least 
> twice in THRIFT, and other projects (Cassandra, for example) have also had to 
> reimplement these classes as a workaround. The reported THRIFT issue would 
> essentially fix it permanently.
>     
>     However, even if that fix is addressed in a future version of Thrift, it 
> will not help us today, and since Thrift 0.9.1 is the latest release, we need 
> this patch today, to integrate with downstream packaging which use Thrift 
> 0.9.1.
> 
> Sean Busbey wrote:
>     please either
>     
>     1) link to the THIRFT jira in the javadoc
>     
>     or
>     
>     2) change the kind of relationship on ACCUMULO-1691. "Depends on" reads 
> to me like we won'd do the upgrade until the upstream ticket is fixed. I'd be 
> fine with that, but it would presumably obviate this class. "broken by" maybe?

Sure, I can certainly do both of those things. In any case, I need to update 
this javadoc regardless, because it describes replacing only the 
AsyncFrameBuffer, and I replace both.


- Christopher


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/21043/#review42607
-----------------------------------------------------------


On May 2, 2014, 7:15 p.m., Christopher Tubbs wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/21043/
> -----------------------------------------------------------
> 
> (Updated May 2, 2014, 7:15 p.m.)
> 
> 
> Review request for accumulo, Sean Busbey and kturner.
> 
> 
> Bugs: ACCUMULO-1691
>     https://issues.apache.org/jira/browse/ACCUMULO-1691
> 
> 
> Repository: accumulo
> 
> 
> Description
> -------
> 
> Updates Thrift dependency to 0.9.1 with a hack to access a needed protected 
> field.
> 
> 
> Diffs
> -----
> 
>   pom.xml 43aa5fb 
>   
> server/base/src/main/java/org/apache/accumulo/server/util/CustomNonBlockingServer.java
>  PRE-CREATION 
>   server/base/src/main/java/org/apache/accumulo/server/util/TServerUtils.java 
> 6d9e4c7 
> 
> Diff: https://reviews.apache.org/r/21043/diff/
> 
> 
> Testing
> -------
> 
> mvn clean verify -Psunny
> 
> 
> Thanks,
> 
> Christopher Tubbs
> 
>

Reply via email to