File.resume() docx file busboy finish event

[AdSegura]

Hello!

In order to discard not allowed extensions I use file.resume() to get off not allowed files and let start with another or finish the form upload.

All went smooth until some type of .docx knock my door.

The problem is when we upload this type of file, Busboy never gets finish signal and browser connection hangs until time out, with other not allowed files they get rejected, Busboy knows when the upload finished.

$> node -v
    v10.15.0

Busboy version:

    "busboy": "^0.3.0",

To reproduce this behaviour:

$> curl https://transfer.sh/K1mPC/f2.docx -o f2.docx

$> file -bi 'f2.docx' 
    application/zip; charset=binary

$> node server.bug.js 

$> curl -i -X POST -H "Content-Type: multipart/form-data" -F "data=@f2.docx" http://localhost:3000/busboy

    HTTP/1.1 100 Continue

    curl: (56) Recv failure: Connection reinitialized by the remote machine

-Server Output:

    readFirstBytes...
    File [data]: filename: f2.docx, encoding: 7bit, mimetype: application/octet-stream
    readFirstBytes...
    Rejected file of type docx
    File ON ERROR [f2.docx] ERROR ->  EXT NOT ALLOWED
    File ON END [f2.docx] Finished
    Reached the end, but did not read anything.

I can upload other doc or docx files and they get rejected and busboy.on('finish') gets executed:

$> curl https://transfer.sh/pyFD3/f1.docx -o f1.docx

$> file -bi 'f1.docx' 
    application/octet-stream; charset=binary

$> curl -i -X POST -H "Content-Type: multipart/form-data" -F "data=@f1.docx" http://localhost:3000/busboy

    HTTP/1.1 100 Continue

    HTTP/1.1 200 OK
    Date: Tue, 22 Jan 2019 07:35:28 GMT
    Connection: keep-alive
    Content-Length: 18

    Done parsing form!

-Server Output:

    readFirstBytes...
    File [data]: filename: f1.docx, encoding: 7bit, mimetype: application/octet-stream
    readFirstBytes...
    Rejected file of type docx
    File ON ERROR [f1.docx] ERROR ->  EXT NOT ALLOWED
    File ON END [f1.docx] Finished
    Reached the end, but did not read anything.
    Done parsing form!

server.js

const http = require('http');
const fs = require('fs');
const path = require('path');
const Busboy = require('busboy');
const fileType = require('file-type');
const port = 3000;
const host = "localhost";

/**
 * Create Server
 */
const server = http.createServer((req, res) => {
   const url = req.url;
   if (url === '/') {
      return renderForm(res);
   } else if (url === '/busboy' && req.method === 'POST') {
      return uploadFiles(req, res);
   } else {
      return notFound(res);
   }
})

server.listen(port, host, () => {
   console.log(`Server running at http://${host}:${port}/`);
 });


/**
 * Upload Files
 * 
 * @param {*} req 
 * @param {*} res 
 */
const uploadFiles = (req, res) => {
   const opt = {
      dest: path.join(__dirname, '.'),
      limits: {
         fileSize: 1024 * 1024,
         files: 2
      }
   };

   var busboy = new Busboy({
      headers: req.headers,
      opt: opt
   });


   busboy.on('file', (fieldname, file, filename, encoding, mimetype) => {

      function readFirstBytes() {
         console.log("readFirstBytes...")
         var chunk = file.read(fileType.minimumBytes);
         if (!chunk) return file.once('readable', readFirstBytes);
         const type = fileType(chunk);
         let ext = "foo";
         if (type) ext = type.ext
         file.unshift(chunk);

         if (ext === 'jpeg' || ext === 'jpg' || ext === 'png') {
            console.log("checkExtention allow ---> ", ext)
            file.pipe(fs.createWriteStream('./' + filename));
         } else {
            console.log('Rejected file of type ' + ext);
            file.resume().on('end', () => {
               console.log('Reached the end, but did not read anything.');
            }); // Drain file stream to continue processing form
            file.emit('error', 'EXT NOT ALLOWED')
            file.emit('close')
            file.emit('end')
         }
      }

      readFirstBytes();

      console.log('File [' + fieldname + ']: filename: ' + filename + ', encoding: ' + encoding + ', mimetype: ' + mimetype);

      file.on('error', function (e) {
         console.log('File ON ERROR [' + filename + '] ERROR -> ', e);
      }).on('end', function () {
         console.log('File ON END [' + filename + '] Finished');
      }).on('limit', () => {
         console.log(`bytes limit ${opt.limits.fileSize} exceeded on File: ${filename}`);
         file.resume(); //fires finish
      });

   }).on('filesLimit', () => {
      console.log('MAX FILES REACHED, LIMIT IS: ', opt.limits.files)
   }).on('finish', (f) => {
      console.log('Done parsing form!');
      return res.end('Done parsing form!');
   });

   req.pipe(busboy);

}

/**
 * Not Found Res
 * 
 * @param {*} res 
 */
const notFound = (res) => {
   res.statusCode = 404;
   res.setHeader('Content-Type', 'text/html');
   res.end('<h1>404 NOT FOUND</h1>');
}

/**
 * Render Upload Form
 * 
 * @param {*} res 
 */
const renderForm = (res) => {
   res.write('<h1>Upload Files<h1>'); 
   res.end(
      '<form action="/busboy" enctype="multipart/form-data" method="post">' +
      '<input type="text" name="title"><br>' +
      '<input type="file" name="upload" multiple="multiple"><br>' +
      '<input type="submit" value="Upload">' +
      '</form>'
   );
}

Thanks for reading.

[AdSegura]

Hi.

Think I just solve it, still doing some tests.

Calling file.resume() on next tick do the trick.

function readFirstBytes() {
         console.log("readFirstBytes...")
         var chunk = file.read(fileType.minimumBytes);
         if (!chunk) return file.once('readable', readFirstBytes);
         const type = fileType(chunk);
         let ext = "foo";
         if (type) ext = type.ext
         file.unshift(chunk);

         if (ext === 'jpeg' || ext === 'jpg' || ext === 'png') {
            console.log("checkExtention allow ---> ", ext)
            file.pipe(fs.createWriteStream('./' + filename));
         } else {
            console.log('Rejected file of type ' + ext);
            process.nextTick(function() {
                file.resume().on('end', () => {
                    console.log('Reached the end, but did not read anything.');
                });                                    
            });
      }

[mscdex]

I think this is a bug in node v10.x. node v8.x and node master currently work as you're expecting.

[mscdex]

Upstream bug report: nodejs/node#25642

[AdSegura]

Thanks for your responses @mscdex appreciate it.

[mscdex]

Alright as noted in the upstream issue, this should've been fixed in node v10.15.1.