Service Worker Lifecycle to improve [via Algorithmic Modelling Community Group]

I m testing the possibilities offered by service worker for now and for the
future.

I m trying to create a websocket proxy with a service worker. Pratically it
keeps up the connection even if normally when you change page or you close the
page the connection was destroyed.It is a basic behaviour for a real web
application using websocket. Conceptually is a error to not foresee that
websocket life is indipendent from life of a page.

Websocket is simply a communication that can remain active for long time or can
be actived in a moment not related to the browser navigation.

The code developed  works partially , so the browser is using now a single
connection  for multiple clients but:
1) i have no event "stopping/starting" in service worker for understand when to
close/open the connection in intelligent way. In general i think it is missing
onstart/onstop events.
2) if the service worker is stopped and i trying to send a event from another
client (different browser ) the service worker doesn't wake up. Web socket is
not only a object but a connection ... so it might have a particolar management
inside service worker. I could register this "alive objects" inside the service
worker the have the ability to wake up a service worker so i can solve this
problem and i could extend it also for pushing (Event sent messages). The push
notification model seams too much trial for the new challenges of the future.

I thing they could be added to w3c specifications.



----------

This post sent on Algorithmic Modelling Community Group



'Service Worker Lifecycle to improve'

https://www.w3.org/community/model/2016/05/26/service-worker-lifecycle-to-improve/



Learn more about the Algorithmic Modelling Community Group: 

https://www.w3.org/community/model

Received on Thursday, 26 May 2016 15:19:51 UTC