Notification storage should use full origin when persisting notifications

NEW
Unassigned

Status

()

Core
DOM
3 years ago
a year ago

People

(Reporter: nsm, Unassigned, Mentored)

Tracking

Trunk
Points:
---

Firefox Tracking Flags

(firefox41 affected)

Details

(Whiteboard: [lang=C++,js])

The Notification storage service uses origins for webpages and manifest urls for apps. Now that the OriginAttributes are encoded in the origin, we should always use that.
Mentor: nsm.nikhil
Whiteboard: [lang=C++,js]

Comment 1

3 years ago
I would like to work on this bug
Michael,

If the user were to update the browser/phone is it acceptable to lose existing notifications or will we need code to perform a clean migration?

Jinank,

I'd like to hear what is your plan to implement this based on your understanding of the code. I'm happy to assist. This would be a good way to implement a change from planning to shipping.
Flags: needinfo?(mhenretty)
Flags: needinfo?(jinank94)
(In reply to Nikhil Marathe [:nsm] (needinfo? please) from comment #2)
> Michael,
> 
> If the user were to update the browser/phone is it acceptable to lose
> existing notifications or will we need code to perform a clean migration?

We will need to make sure we don't lose APP notifications (ie. ones tied to a manifestURL), but we don't care about browser notifications. Not sure if this will require migration logic, or just supporting the old and new DB format.

A little background here: since bug 1159128 we only store _app_ notifications, and not browser notifications. Notification.get was removed from the spec, so we only keep the database around for resending notifications in b2g right now. Later we will need to revive the Notification DB to support persistent notifications tied to a service worker. Perhaps this bug makes more sense to be completed as part of that work.
Flags: needinfo?(mhenretty)

Comment 4

2 years ago
For getting the better understanding about the problem could you please tell which files would be relevant to this bug?
Flags: needinfo?(jinank94)
You need to log in before you can comment on or make changes to this bug.