slugboot

aggressively cache at the application layer for offline-only webapps

With this module serving as a kind of web bios, you can create identical "slug" domains (with https certs) that can be flashed with any application. The application payload is stored in indexedDB.

This way, you can load webapps from p2p networks over webrtc, bluetooth LE, or acoustic couplers. The webapps stay saved on your computer and you can open them on a slug domain whenever you wish.

security

slugboot uses trust on first use to initialize a generic webapp container that cannot be updated except through an application's own direct intervention.

Unfortunately, until ServiceWorker version 2 ships with a Service-Worker-Max-Age header, service worker Max-Age is capped to 24 hours. This means the original domain that loaded the slugboot worker could later ship a malicious version of /worker.js.

To mitigate this attack for now, users should add a record in /etc/hosts for slugbooted domains that points back to localhost. For example:

127.0.0.1    slug.example.com

The request or certificate check will fail for requests to /worker.js, so the old worker will continue to run as before.

example

create a main.js file:

var slug = require('slugboot')('/worker.js')
slug.version(function (err, v) {
  if (err) return console.error(err)
  if (v) return console.log('already uploaded a version')

  slug.fetch('/')
  slug.fetch('/bundle.js')
  slug.fetch('/worker.js')
  slug.put('/hello.txt', 'HI THERE')
  slug.commit(function (err) {
    if (err) return console.error(err)
    else console.log('uploaded version')
  })
})

window.slug = slug

set up public/ with the slugboot worker.js file:

$ npm i -g browserify ecstatic
$ mkdir public
$ echo '<script src="bundle.js"></script>' > public/index.html
$ cp `node -pe "require.resolve('slugboot/worker.js')"` public/
$ browserify main.js > public/bundle.js
$ ecstatic -H 'Service-Worker-Max-Age: 3153600000000' -p 44000 public/

Now open http://localhost:44000 and poke around with the slug instance on the REPL.

api

var slugboot = require('slugboot')

var slug = slugboot(workerURL)

Create a new slugboot instance from a workerURL path string to the slugboot service worker code.

slug.put(path, body, cb)

Stage a string body to be saved at the path string path.

cb(err) fires with any errors.

slug.fetch(src, dst, cb)

Request the document at the url string src and save the result to the path at dst.

If dst is not given, it will use the pathname of src.

cb(err) fires with any errors.

slug.copy(src, dst, cb)

Copy the file path src from the current active version to dst in the staging version.

cb(err) fires with any errors.

slug.commit(cb)

Commit any staged changes and increment the version.

cb(err) fires with any errors.

slug.version(cb)

Query for the integer version as cb(err, version).

Before flashing, the version will be undefined.

The first version after a commit is version 1.

install

npm install slugboot

license

BSD