-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/18584/#review36571
-----------------------------------------------------------

Ship it!


Ship It!

- Ben Mahler


On March 7, 2014, 5:20 p.m., Jie Yu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/18584/
> -----------------------------------------------------------
> 
> (Updated March 7, 2014, 5:20 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Ben Mahler.
> 
> 
> Bugs: MESOS-984
>     https://issues.apache.org/jira/browse/MESOS-984
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> Previously, we use onAny to connect most of the steps. As a result, we see a 
> lot of the error checking code like the following:
> 
> void checked(const Future<Metadata::Status>& future)  
> {
>   if (!future.isReady()) {
>     promise.fail(
>         future.isFailed() ?
>         "Failed to get replica status: " + future.failure() :
>         "Not expecting discarded future");
>     terminate(self());
>     return;
>   }
>   ...
> }
> 
> Another drawback is: discarding becomes difficult and hard to reason.
> 
> Now, I cleaned up the code to use the continuation style (i.e., chaining 
> using Future.then). The code becomes much more clear now!
> 
> Also, I factored out the recover protocol part from the log recover process 
> to make the log recover process less cumbersome.
> 
> This patch does not change any semantics, but is for MESOS-984 (log auto 
> initialization for registrar). Will follow up with a patch to integrate the 
> log auto initialization based on this patch.
> 
> 
> Diffs
> -----
> 
>   src/log/recover.cpp e611a4e 
> 
> Diff: https://reviews.apache.org/r/18584/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Jie Yu
> 
>

Reply via email to