mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Anand Mazumdar <an...@apache.org>
Subject Re: Review Request 48094: Implemented GET_ROLES Call in v1 master API.
Date Fri, 17 Jun 2016 23:36:11 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48094/#review138342
-----------------------------------------------------------



Just a few more minor comments. Looks pretty good otherwise!


include/mesos/master/master.proto (lines 286 - 287)
<https://reviews.apache.org/r/48094/#comment203503>

    The `GetRoles` message has much more information than this. Let's be a bit more explicit.
Also, the previous message had some typos.
    
    How about:
    ```cpp
    // Provides information about every role that is on the role whitelist (if enabled), has
one or more registered frameworks or has a non-default weight or quota.
    ```



include/mesos/mesos.proto (line 2015)
<https://reviews.apache.org/r/48094/#comment203515>

    hmm, thinking about it more, we can just do:
    
    `repeated FrameworkID frameworks` now that we have the well define protobuf object instead
of returning a `string`.



include/mesos/v1/master/master.proto (lines 286 - 287)
<https://reviews.apache.org/r/48094/#comment203504>

    Ditto as the unversioned proto comment.



src/master/http.cpp (line 1417)
<https://reviews.apache.org/r/48094/#comment203510>

    hmm, can we include some information as to why we ended up duplicating most of the logic
of `_roles` for posterity?
    
    How about:
    ```cpp
    This duplicates the functionality offered by `roles()`. This was necessary as the JSON
object returned by `roles()` was not specified in a formal way i.e. via a corresponding protobuf
object and would have been very hard to convert back into a `Resource` object.
    ```



src/master/http.cpp (line 1430)
<https://reviews.apache.org/r/48094/#comment203511>

    s/getRole/role
    
    Also move this down inside the `foreach` loop on L1449



src/master/http.cpp (lines 1469 - 1472)
<https://reviews.apache.org/r/48094/#comment203518>

    Can't you just do:
    
    ```cpp
    getRole.mutable_resources()->CopyFrom(role.get()->resources);
    ```



src/tests/api_tests.cpp (line 317)
<https://reviews.apache.org/r/48094/#comment203521>

    Can you also include an assert for the expected number of roles in the response i.e. 3?
    
    ```cpp
    ASSERT_EQ(3, v1Response->get_roles().roles().size());
    ```



src/tests/api_tests.cpp (lines 319 - 322)
<https://reviews.apache.org/r/48094/#comment203527>

    Can you directly compare the `Resource` object itself instead of comparing some resource
atrributes?


- Anand Mazumdar


On June 17, 2016, 8:30 a.m., Abhishek Dasgupta wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48094/
> -----------------------------------------------------------
> 
> (Updated June 17, 2016, 8:30 a.m.)
> 
> 
> Review request for mesos, Anand Mazumdar and Vinod Kone.
> 
> 
> Bugs: MESOS-5494
>     https://issues.apache.org/jira/browse/MESOS-5494
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Implemented GET_ROLES Call in v1 master API.
> 
> 
> Diffs
> -----
> 
>   include/mesos/master/master.proto 3b85b8130f1a292f3574214fcf7761c891f57f22 
>   include/mesos/mesos.proto d3cfdd0a583bbd0f5ec11c87ea29e649a97450c5 
>   include/mesos/v1/master/master.proto 5a6bbeb0f9b57d03803e366df32e04e9598e00ed 
>   include/mesos/v1/mesos.proto dca0c2c8a468345cb18b4ced889fc3db0324ca1d 
>   src/master/http.cpp 1b74211acafc15cc0bcb24545ca4c2bacd79cb2d 
>   src/master/master.hpp 72c60ef74ce57119a97cf8305182340a13c58c42 
>   src/tests/api_tests.cpp 03e46cc5e0af8250ba36281e1293a1dc89f8e266 
> 
> Diff: https://reviews.apache.org/r/48094/diff/
> 
> 
> Testing
> -------
> 
> On Ubuntu 16.04:
> 
> sudo GTEST_FILTER="*MasterAPITest.*" make -j2 check
> 
> 
> Thanks,
> 
> Abhishek Dasgupta
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message