Adds a new controlledBySW method

[jeffposnick]

R: @gauntface

Fixes #53

[gauntface]

All good - just please add a comment to the skipWaiting service worker please.

Will publish a release once it's merged.

[gauntface]

Could you add a comment here just along the lines of:

"This service worker does not use event.waitUntil() on purpose. This has ambiguity over implementation across browsers (i.e. whether it's needed to be used with skipWaiting and clients.claim()). This service worker is used to test controlledBySw() - PLEASE DO NOT CHANGE."

Without this I can see someone eventually sneaking in event.waitUntil() (i.e. myself) and then it's undermines the purpose of this new method.

[jeffposnick]

I'm not sure what the ambiguity is, though? AFAIK there are no browsers that ever required an event.waitUntil() wrapper around self.skipWaiting() or self.clients.claim().

There are old examples that used event.waitUntil() (including many I wrote), but it turns out it was never actually necessary.

[jeffposnick]

(I added a comment anyway!)

[gauntface]

But we've seen the difference in behavior - so it's not "required" it does affect the end behavior.

[jeffposnick]

PTAL at the revised comment.

[gauntface]

yeah that works.

[jeffposnick]

I've made eslint happier with some of the latest commits, but the actual tests are failing now with

WebDriverError: Unable to parse new session response: {"value": {"sessionId":"e8da8745-f4ac-46c5-ba1f-de5b36a55a4d","value":{"acceptInsecureCerts":false,"browserName":"firefox","browserVersion":"55.0a1","moz:accessibilityChecks":false,"moz:processID":7046,"moz:profile":"/tmp/rust_mozprofile.RHnTV5h5iha9","pageLoadStrategy":"normal","platformName":"linux","platformVersion":"4.8.12-040812-generic","rotatable":false,"specificationLevel":0,"timeouts":{"implicit":0,"pageLoad":300000,"script":30000}}}}

I'm going to add in another commit that bumps all the devDependenciesand see if that helps, unless you have other ideas, @gauntface.

[jeffposnick]

Updating the dependencies is pretty involved, unfortunately, so I'll tackle it in a separate PR that will need to be merged first.

[jeffposnick]

This PR is blocked on #55, as I haven't had any luck getting the test suite to run properly in Firefox 😦

[gauntface]

@jeffposnick if you can switch back to let instead of disabling the lint check, I'll merge the PR and will tackle the tests in a follow up PR if you like.

[jeffposnick]

Oops—my forked copy was out of date with the upstream master, and I didn't realize that the rest of that file had switched from var to let. I couldn't figure out why linting was unhappy with just that new line... it's switched to let now.

[gauntface]

@jeffposnick Tank you :) Will make sure lint is at least happy + Chrome and then will merge. Sorry for the faff.