mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vinod Kone <vinodk...@gmail.com>
Subject Re: Review Request 54497: Added a test for the 'ATTACH_CONTAINER_INPUT' call.
Date Wed, 07 Dec 2016 22:05:55 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/54497/#review158388
-----------------------------------------------------------




src/tests/api_tests.cpp (line 3990)
<https://reviews.apache.org/r/54497/#comment229147>

    why is this an Option?



src/tests/api_tests.cpp (line 4065)
<https://reviews.apache.org/r/54497/#comment229149>

    s/message/processIO/



src/tests/api_tests.cpp (line 4084)
<https://reviews.apache.org/r/54497/#comment229150>

    s/message/processIO/



src/tests/api_tests.cpp (lines 4094 - 4118)
<https://reviews.apache.org/r/54497/#comment229151>

    can't we use http::post() here? would be nice to add support to send streaming request
to that function.



src/tests/api_tests.cpp (lines 4121 - 4146)
<https://reviews.apache.org/r/54497/#comment229177>

    so this looks different from how we read piped records in io switchboard tests (and also
my review).
    
    this is how it looks:
    ```
      http::Pipe::Reader reader = response->reader.get();
    
      auto deserializer =
        lambda::bind(deserialize<v1::agent::ProcessIO>, contentType, lambda::_1);
    
      Reader<v1::agent::ProcessIO> decoder(
          ::recordio::Decoder<v1::agent::ProcessIO>(deserializer), reader);
    
      string stdoutReceived = "";
      string stderrReceived = "";
    
      while (true) {
        Future<Result<v1::agent::ProcessIO>> _message = decoder.read();
        AWAIT_READY(_message);
    
        if (_message->isNone()) {
          break;
        }
    
        ASSERT_SOME(_message.get());
    
        v1::agent::ProcessIO message = _message->get();
    
        ASSERT_EQ(v1::agent::ProcessIO::DATA, message.type());
    
        ASSERT_TRUE(message.data().type() == v1::agent::ProcessIO::Data::STDOUT ||
                    message.data().type() == v1::agent::ProcessIO::Data::STDERR);
    
        if (message.data().type() == v1::agent::ProcessIO::Data::STDOUT) {
          stdoutReceived += message.data().data();
        } else if (message.data().type() == v1::agent::ProcessIO::Data::STDERR) {
          stderrReceived += message.data().data();
        }
      }
    
      EXPECT_EQ(output, stdoutReceived);
      EXPECT_EQ(error, stderrReceived);
    
    ```
    
    maybe we should write an helper so that we can standardize and de-duplicate. how about
    
    ```
    // Reads ProcessIO::Data records for the pipe `reader` until EOF is reached and returns

    // merged stdout and stderr as a tuple. Note: It ignores any ProcessIO::Control records.
    Try<tuple<string, string>> getProcessIOData(http::Pipe::Reader reader)
    {
    
    }
    
    ```


- Vinod Kone


On Dec. 7, 2016, 8:15 p.m., Anand Mazumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/54497/
> -----------------------------------------------------------
> 
> (Updated Dec. 7, 2016, 8:15 p.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added a test for the 'ATTACH_CONTAINER_INPUT' call.
> 
> 
> Diffs
> -----
> 
>   src/tests/api_tests.cpp 6cd1f83a47731ee8c4fb1f022bf585b271b3b966 
> 
> Diff: https://reviews.apache.org/r/54497/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Anand Mazumdar
> 
>


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