Sometimes the message scheduled & message sent notifications are
received out of order.
Now we use a single channel for both so order is maintained.
This was causing the tests to be flaky and it might have happened in
production as well.
This commit fixes one source of flakyness in the tests, which was an
actual bug.
If 1 device is registering with a push notification server, if there's
another device with the same public key, both would mark themselves as
registered, while maybe only one has been actually registered.
To fix this, we keep track of the request ids we send (in memory for
now), and only mark it as registered if the request was originating on
this device.
This commit introduces the following changes:
- `local-notifications` require as body an interface complying with
`json.Marshaler`
- removed unmarshaling of `Notifications` as not used (we only Marshal
notifications)
- `protocol/messenger.go` creates directly a `Notification` instead of
having an intermediate format
- add community notifications on request to join
- move parsing of text in status-go for notifications
One of the issues we noticed is that the partitioned topic
in push notification is heavy in traffic, as any user using a particular
mailserver will use that partitioned topic to register for PNs.
This commit moves from the partitioned topic to the personal topic of
the PN server, so it does not clash with other users that might happen
to have the same partitioned topic as the mailserver, resulting in long
sync times.
Another issue that will need to be addressed separately is that once you
send a message to a topic, because of the way how waku/whisper works,
you will have to register to that topic, meaning that you will receive
that data. Currently waku does not support unsubscribing from a topic
without logging in and out, so that needs also to be addressed.
In some instances the retry mechanism would get into a busy loop.
That's due to the fact that we would fetch some non-retriable
notifications but not act on them.
This commit fixes the issue by filtering them from the database query
and making sure that we at least wait 1 second.
We were checking for the wrong error kind when pulling messages from the
database, which resulted in the code not retrying to pull the message,
giving flaky tests / race condition (that's present in production as
well)
Currently messenger has no notion of being online.
This might cause a problem as we retry to register with a push
notification server even if not connected to any peer, which will
inevitably fail.
This commit adds a method `handleConnectionChange` that will be called
every time the connection change state.
Ibrahem Khalil
Andrea Maria Piana
Volodymyr Kozieiev