Abort trap: 6

[claudio-destro]

Hi guys

Chatted with @reconbot about the following error I got in my code:

Assertion failed: (loop->watchers[w->fd] == w), function uv__io_stop, file ../deps/uv/src/unix/core.c, line 864.
Abort trap: 6

it seems to be random... I got it in my library flashmagic.js branch develop.

please run cli.ts with subcommand "ping" but I think you will need an NXP LPC hardware...

# ./bin/flashmagic.js -P <your_port> ping

I get the error after attempting to close and reopen a given serial port with different baud rate (see last lines in src/Handshake.ts): 9600 -> 115200.

---

Attempting to reproduce the issue I wrote the following code in plain javascript: it fails fast sometimes with "Error: Serialport not open" (-> uncomment on('data')), sometimes with "Error: Cannot open /dev/tty.usbserial-FTBTBAAW".

It seems like the close/open functions do not work as expected.

var com = require('serialport');

var path = process.argv[2]

var serialport = openCom(path, 9600);
var count = 1;

(function loop() {
  // serialport.on('data', function(data) {
  //   console.log(data);
  // });
  serialport.open(function(error) {
    if (error) { console.error(error); throw error;}
    serialport.write('?', function(error) {
      if (error) { console.error(error); throw error;}
      serialport.drain(function(error) {
        if (error) { console.error(error); throw error;}
        serialport.close(function(error) {
          if (error) { console.error(error); throw error;}
          console.log('closed ' + count++);
          serialport = openCom(path, 115200);
          process.nextTick(loop);
        }); // close
      }); // drain
    }); // write
  });
})();

function openCom(path, baud) {
  return new com.SerialPort(path, {
      baudRate: baud,
      stopBits: 1,
      parity: 'none',
      parser: com.parsers.readline('\r\n')
    }, false); // open later
}

/*
$ uname -a
Darwin iMac-di-Claudio.lan 15.2.0 Darwin Kernel Version 15.2.0: Fri Nov 13 19:56:56 PST 2015; root:xnu-3248.20.55~2/RELEASE_X86_64 x86_64
$ node -v
v5.4.0
*/

[claudio-destro]

See this
nodejs/node#3604

[reconbot]

Can you confirm your platform? I think this is related to #670

[claudio-destro]

Confirmed. I think it is related to open/close port too: actually I have to use only one port at a time.

Inviato da iPad

Il giorno 27 mar 2016, alle ore 06:16, Francis Gulotta [email protected] ha scritto:

Can you confirm your platform? I think this is related to #670

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub

[reconbot]

I meant, is this Linux on an embedded device, and if so, which one?

[claudio-destro]

Nope

As written in the code, it is OS X with node 5.4

Claudio

Il giorno 27 mar 2016, alle ore 17:11, Francis Gulotta [email protected] ha scritto:

I meant, is this Linux on an embedded device, and if so, which one?

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub

[reconbot]

@claudio-destro We fixed a ton of bugs in our latest beta specifically ones addressing this issue on osx.

#733 Please try it out [email protected] and report back!

[claudio-destro]

Ok I will try it out!

Inviato da iPhone

Il giorno 06 apr 2016, alle ore 19:20, Francis Gulotta [email protected] ha scritto:

@claudio-destro We fixed a ton of bugs in our latest beta specifically ones addressing this issue on osx.

#733 Please try it out [email protected] and report back!

―
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub

[reconbot]

Did it work out? We've released [email protected].

[claudio-destro]

I didn't have enough time... Sorry

I did a little test yesterday night and seemed working. I will try the new release.

Claudio

Inviato da iPhone

Il giorno 11 apr 2016, alle ore 03:22, Francis Gulotta [email protected] ha scritto:

Did it work out? We've released [email protected].

―
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub

[reconbot]

I'm going to close this issue due to it's age, but if you'd like to continue with it feel free to reopen.

[reconbot]

Can you show me the crashing code? And can you tell me what version you're
running.

---

Francis Gulotta
[email protected]

On Wed, Nov 9, 2016 at 3:15 PM, Luc DeTellis [email protected]
wrote:

I was getting the "Abort trap: 6" crash. After some experimenting I found
that I was sending writes one immediately after another, which was
apparently too fast for serialport to handle.

Here's the setTimeout I had to add:

serial.write(output + '\n', function(err) {
if (err) return log.error('serialport', err);
setTimeout(function() {
if (cue.length === 0) {
canSend = true;
} else {
write(cue.shift());
}
}, 0);
});

—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub
#660 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AABlbq837CmEB5dEkI_3eRqXeZgDp84Lks5q8in_gaJpZM4HK9y2
.

[lucdetellis]

@reconbot Sorry I deleted my original comment, I found a better solution than using a setTimeout.

Here's my original code if you're interested:

var write = function(output) {
  if (serial.isOpen) {
    serial.write(output + '\n', function(err) {
      if (err) return log.error('serialport', err);
      if (cue.length === 0) {
        canSend = true;
      } else {
        write(cue.shift());
      }
    });
  }
};

var send = function(output) {
  if (canSend === true) {
    canSend = false;
    write(output);
  } else {
    cue.push(output);
  }
};

// EDIT: Only 2 writes were enough to trigger "Abort trap: 6"
send('1,0,0,0');
send('exec');

I fixed the "Abort trap: 6" error by adding serial.drain() before each write:

var write = function(output) {
  serial.write(output + '\n', function(err) {
    if (err) return log.error('serialport', err);
    serial.drain(function(err) {
      if (err) return log.error('serialport', err);
      if (cue.length === 0) {
        canSend = true;
      } else {
        write(cue.shift());
      }
    });
  });
};

[reconbot]

There isn't an upper limit on how many writes you can do in sequence, so
I'd like to get to the bottom of this.

---

Francis Gulotta
[email protected]

On Wed, Nov 9, 2016 at 4:00 PM, Luc DeTellis [email protected]
wrote:

@reconbot https://github.com/reconbot Sorry I deleted my original
comment, I found a better solution than using a setTimeout.

Here's my original code if you're interested:

var write = function(output) {
if (serial.isOpen) {
serial.write(output + '\n', function(err) {
if (err) return log.error('serialport', err);
if (cue.length === 0) {
canSend = true;
} else {
write(cue.shift());
}
});
}
};

var send = function(output) {
if (canSend === true) {
canSend = false;
write(output);
} else {
cue.push(output);
}
};

send('a lot of things at once');
...
send('more things');

I fixed the "Abort trap: 6" error by adding serial.drain() before each
write:

var write = function(output) {
serial.write(output + '\n', function(err) {
if (err) return log.error('serialport', err);
serial.drain(function(err) {
if (err) return log.error('serialport', err);
if (cue.length === 0) {
canSend = true;
} else {
write(cue.shift());
}
});
});
};

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#660 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AABlblEXHQ8CYIAJVZDMffnj0VmdL9U3ks5q8jSFgaJpZM4HK9y2
.