Skip to content

Improves node.js stacktraces and makes it easier to handle errors

License

Notifications You must be signed in to change notification settings

Pita/async-stacktrace

Repository files navigation

Why classic stacktraces are not very helpful when dealing with async functions

Look at this example. one calls two, two calls three, and three calls four. All functions call the given callback asynchronous. four calls the callback with an error. three and two passes the error to their callback function and stop executing with return. one finally throws it

function one()
{
   two(function(err){
     if(err){
       throw err;
     }
   
     console.log("two finished");
   });
}

function two(callback)
{
  setTimeout(function () { 
    three(function(err)
    {
      if(err) {
        callback(err);
        return;
      }
      
      console.log("three finished");
      callback();
    });
  }, 0);
}

function three(callback)
{
  setTimeout(function () { 
    four(function(err)
    {
      if(err) {
        callback(err);
        return;
      } 
      
      console.log("four finished");
      callback();
    });
  }, 0);
}

function four(callback)
{
  setTimeout(function(){
    callback(new Error());
  }, 0);
}

one();

When you execute it, you will get this:

$ node example_without.js 

/home/pita/Code/async-stacktrace/example_without.js:5
       throw err;
       ^
Error
    at Timer.callback (/home/pita/Code/async-stacktrace/example_without.js:47:14)

The problems here are:

  • You can see that the error happend in four, but you can't see from where four was called. The context gets lost
  • You write the same 4 lines over and over again, just to handle errors

The solution

Lets replace this code in two and three

if(err) {
  callback(err);
  return;
}

with

if(ERR(err, callback)) return;

and replace this code in one

if(err){
  throw err;
}

with

ERR(err);

This is how it looks like now:

var ERR = require("async-stacktrace");

function one()
{
   two(function(err){
     ERR(err);
   
     console.log("two finished");
   });
}

function two(callback)
{
  setTimeout(function () { 
    three(function(err)
    {
      if(ERR(err, callback)) return;
      
      console.log("three finished");
      callback();
    });
  }, 0);
}

function three(callback)
{
  setTimeout(function () { 
    four(function(err)
    {
      if(ERR(err, callback)) return;
      
      console.log("four finished");
      callback();
    });
  }, 0);
}

function four(callback)
{
  setTimeout(function(){
    callback(new Error());
  }, 0);
}

one();

When you execute it, you will get this:

$ node example.js 

/home/pita/Code/async-stacktrace/ERR.js:57
      throw err;
      ^
Async Stacktrace:
    at /home/pita/Code/async-stacktrace/example.js:6:6
    at /home/pita/Code/async-stacktrace/example.js:17:10
    at /home/pita/Code/async-stacktrace/example.js:30:10

Error
    at Timer.callback (/home/pita/Code/async-stacktrace/example.js:41:14)

What is new?

The "Async Stacktrace" shows you where this error was caught and passed to the next callback. This allows you to see from where four was called. You also have less code to write

npm

npm install async-stacktrace

Usage

This is how you require the ERR function

var ERR = require("async-stacktrace");

The parameters of ERR() are:

  1. err The error object (can be a string that describes the error too)
  2. callback (optional) If the callback is set and an error is passed, it will call the callback with the modified stacktrace. Else it will throw the error

The return value is true if there is an error. Else its false

About

Improves node.js stacktraces and makes it easier to handle errors

Resources

License

Stars

Watchers

Forks

Packages

No packages published