chainr

flow control library with a chaining interface, inspired by node-seq by substack

npm install chainr
13 downloads in the last month

Chainr

Build Status

NPM

NPM

Simple flow control library inspired by seq by substack.

The interface is quite similar but with a reduced function range. Instead of using the this property to assign a done-callback the traditional node way is used (a callback function as the last parameter).

API

.seq(cb)

.seq(name, cb)

Functions registered via .seq are executed in sequential order. Each cb gets a callback assigned which is used to signal the end of the callback execution. If name is specified, the second argument sent to cb goes to chain.var.

example/seq.js

var chainr = require('../');

var chain = chainr();

chain
    .seq('foo', function (cb) {
        console.log('1');
        setTimeout(cb.bind(null, null, 'bar'), 100);
    })
    .seq(function (cb) {
        console.log('2');
        setTimeout(cb, 100);
    })
    .seq(function (cb) {
        console.log('3');

        console.log(chain.vars);
    });

Output:

1
2
3
{ foo: 'bar' }

.par(cb)

.par(name, cb)

Functions registered via .par are executed in parallel. Each cb gets a callback assigned which is used to signal the end of the callback execution. If name is specified, the second argument sent to cb goes to chain.var.

example/par.js

var chainr = require('../');

var chain = chainr();

chain
    .par(function(cb) {
        setTimeout(function() {
            console.log('1.1');
            cb();
        }, 1000);
    })
    .par('foo', function(cb) {
        setTimeout(function() {
            console.log('1.2');
            cb(null, 'bar');
        }, 400);
    })
    .par(function(cb) {
        setTimeout(function() {
            console.log('1.3');
            cb();
        }, 100);
    })
    .seq(function (cb) {
        console.log(chain.vars);
    });

Output:

1.3
1.2
1.1
{ foo: 'bar' }

.one(cb)

.one(name, cb)

Functions registered via .one are executed in parallel. one acts equally as par with the distinction that it does not wait for all callbacks to execute, one call is sufficient. Each cb gets a callback assigned which is used to signal the end of the callback execution. If name is specified, the second argument sent to cb goes to chain.var. (Notice that not every value might be in there!)

example/one.js

var chainr = require('../');

var order = [];
var chain = chainr();

chain
    .one(function(cb) {
        setTimeout(function() {
            console.log('1.1');
            order.push('1.1');
            cb();
        }, 1000);
    })
    .one('foo', function(cb) {
        setTimeout(function() {
            console.log('1.2');
            order.push('1.2');
            cb(null, 'bar');
        }, 400);
    })
    .one(function(cb) {
        setTimeout(function() {
            console.log('1.3');
            order.push('1.3');
            cb();
        }, 100);
    })
    .seq(function (cb) {
        console.log(order);
    });

Output:

1.3
1.2
1.1
['1.1']

.catch(cb)

When in a sequential or parallel context a callback gets an error object assigned, the execution steps over all pending steps and skips to the next catch block. The catch block than receives the error and an callback that can be used to continue with the execution or to rethrow the error.

example/catch.js

var chainr = require('../');

chainr()
    .seq(function (cb) {
        cb(new Error('foo'));
    })
    .seq(function (cb) {
        console.log('2');
        cb();
    })
    .catch(function (error, cb) {
        console.log(error);
        cb(error);
    })
    .catch(function(error, cb) {
        console.log(error);
        cb();
    })
    .seq(function() {
        console.log('final');
    });

Output:

[Error: foo]
[Error: foo]
final
npm loves you