mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Anand Mazumdar <mazumdar.an...@gmail.com>
Subject Re: Review Request 42181: Added a process ID argument to Slave constructor.
Date Wed, 03 Feb 2016 02:31:55 GMT


> On Feb. 2, 2016, 10:05 p.m., Vinod Kone wrote:
> > src/tests/cluster.cpp, lines 409-430
> > <https://reviews.apache.org/r/42181/diff/1/?file=1193937#file1193937line409>
> >
> >     instead of doing if else blocks can you use a ternary operator for the 1st argument
in the constructor?
> >     
> >     ```
> >     slave.slave = new Slave::Slave(
> >       id.isSome() ? id.get() : process::ID::generate("slave"),
> >       ...
> >       ...);
> >     ```
> >     
> >     infact instead of declaring an overloaded constructor in slave.hpp and chaining
constructors, can you make 'id' an argument with default value?
> >     
> >     ```
> >     Slave(...,
> >           const std::string& id = process::ID::generate("slave"))
> >     ```
> >     
> >     Thoughts?

As per our discussion, modified the existing `Slave` constructor to take an "id" as argument.
Also, cleaned up other places in code to use this new constructor.


- Anand


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42181/#review117492
-----------------------------------------------------------


On Feb. 3, 2016, 2:31 a.m., Anand Mazumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42181/
> -----------------------------------------------------------
> 
> (Updated Feb. 3, 2016, 2:31 a.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Bugs: MESOS-4255
>     https://issues.apache.org/jira/browse/MESOS-4255
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This change modifies the existing `Slave` constructor to take in the process ID string
as an argument. This is necessary for testing recovery of HTTP based executors. (Previously,
every invocation of `StartSlave` used to generate a new ID via `ID::generate`). There was
no process delegate set via `process::initialize` that led to problems for the existing HTTP
based executor to connect back to the slave.
> 
> Also, modified the tests to introduce a new `StartSlave` function that takes this process
ID as argument.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.hpp 632640293f14fb97aea6d7dd2c462c1f604638fb 
>   src/slave/slave.cpp 1f4c8368feb0ce19963577582ce745acfb21aa9f 
>   src/tests/cluster.hpp 576dcb8b7e27d1905425aa0f7cb319c19c17c15c 
>   src/tests/cluster.cpp 1a3a038e78999bca2fc3f08725f4ffaf4290bcb9 
>   src/tests/mesos.hpp c2bae4767ee7372c796bfad44ed1e86db7dd3488 
>   src/tests/mesos.cpp 18d0d8f8037ebc27c87bcb0f1ce9f143e7505ec8 
> 
> Diff: https://reviews.apache.org/r/42181/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Anand Mazumdar
> 
>


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