Solving the subset sum problem question

[Ashilex]

Question About Streaming Behavior in the Node.js Server

Hi,

I'm using the Node.js server script you provided at page 456 of the book, and I have a question regarding how responses are sent to the client.

Observed Behavior

When I run the following curl command:

curl -G http://localhost:8000/subsetSum \
  --data-urlencode "data=[16,19,1,1,-16,9,1,-5,-2,17,-15,-97,19,-16,-4,-5,15]" \
  --data-urlencode "sum=0"

I expect the response to be streamed line by line as each match event is emitted. However, what I observe is that all the Match: lines are printed at once, rather than progressively arriving in real-time.

Code in Question

The server code contains the following logic:

createServer((req, res) => {
  const url = new URL(req.url, 'http://localhost');
  if (url.pathname !== '/subsetSum') {
    res.writeHead(200);
    return res.end('I\'m alive!\n');
  }

  const data = JSON.parse(url.searchParams.get('data'));
  const sum = JSON.parse(url.searchParams.get('sum'));
  res.writeHead(200);

  const subsetSum = new SubsetSum(sum, data);
  
  subsetSum.on('match', match => {
    res.write(`Match: ${JSON.stringify(match)}\n`);
  });

  subsetSum.on('end', () => res.end());

  subsetSum.start();
}).listen(8000, () => console.log('Server started'));

My Understanding & Questions

1. Shouldn't res.write() immediately flush each chunk of data to the client?
   • I was expecting each Match: ... line to be sent as soon as res.write() is called.
2. Could there be buffering happening at some level (Node.js, TCP, or curl)?

thanks for the help

[lmammino]

Hello @Ashilex,

thanks for getting a copy of the book and for submitting this question.

Buffering might happen for a variety of reasons (OS, Node.js version, etc)

One way you can force the buffered data to be flushed on demand is to use the following pattern:

writableStream.cork() // starts to buffer all following writes into memory
writableStream.write('...')
writableStream.uncork() // forces a flush of all the buffered data to the underlying writable resource

So if we change the code as in the following screenshot (note lines 19 and 21), you should now see the expected behaviour:

You can find more info about writable.cork() and writable.uncork() in the official Node.js docs.

I hope this helps 😊