fritter

Transform the AST in order to generate stack traces with no stack trace API necessary.

This module is like stackedy, except that it uses esprima by way of falafel.

examples

giant.js:

var fs = require('fs');
var src = fs.readFileSync(__dirname + '/sources/giant.js', 'utf8');

var fritter = require('fritter');
var fry = fritter(src);
fry.on('error', function (err, c) {
    console.log(String(err));
    c.stack.forEach(function (s) {
        console.log(
            '  in .' + fry.nameOf(s) + '()'
            + ', line ' + s.start.line
        );
    });
});
fry.run();

sources/giant.js:

(function fee () {
    [ 'foe', 'fum' ].forEach(fie);
})();

function fie (s) { smell(s) }
function smell () { english('blood') }

output:

ReferenceError: english is not defined
  in .smell(), line 5
  in .forEach(), line 2
  in .fee(), line 1

methods

var fritter = require('fritter')

var fry = fritter(src, context={}, opts={})

Return a fry object given some optional javascript source src and some context to execute in and populate.

fry emits 'error' events with the error object and the present callstack as arguments when exceptions in src are thrown.

If opts.longStacks is true, turn on longer stack traces for a performance penalty.

fry.include(src, opts={})

Include and wrap some javascript source src. Use opts.filename to associate nodes in the stack with filenames so custom error handlers know what file an exception originated form.

fry.run()

Run the fry.source under fry.context using Function() as a nicer form of eval().

fry.stop()

Stop execution insofar as that is possible.

If callbacks from external APIs fire, the body of those callbacks won't fire.

attributes

fry.source

transformed source to execute with something like vm.runInNewContext()

fry.context

context to execute fry.source with, including randomly-named wrapper functions to catch exceptions and track the call stack

fry.names

names of the randomly-generated wrapper functions

install

With npm do:

npm install fritter

license

MIT