It’s unlikely that any dev “kills” his work after being so committed to it and leaves users, whom he knew almost one by one, with a handful of dust.

Normally, in commercial services, push notifications require support from the backend. The mechanism works approximately like this: the mobile app registers to a third-party provider such as Firebase Cloud Messaging (on Android) and APNS (on Apple Push Notification System) and obtains a device token, it then makes a network call to the server to communicate their device token (and the user ID this is associated to) and then the server dispatches notifications using the user ID to determine the events of interest and the device token to have it delivered via Firebase or APNS depending on the platform. Lemmy instances, unfortunately, don’t do anything of that so client apps have to periodically poll for new events, which is terrible for both the battery and the network usage. Raccoon is no exception of this, if you look in the “Advanced Settings” screen under the “Experiments” section, there is an item named “Check for unread items in background”. This is an attempt to use Android’s WorkManager APIs to perform scheduled periodic operations such as checking for new mentions/replies/messages. Unfortunately, the scheduler does not seem so reliable so after some time the app is in the background the pending work can never get to be executed. It was an attempt and it should be refined.

It seems that they have decided to restart here and create a new community here.

This fork is the new official one. A new community has been created recently to replace the old one too.

We are planning to release a new version, at least on GitHub asap. The announcement is likely to be given in tomorrow’s Q&A session (the monthly AMA).

update: we have decided to continue the project starting from the latest active fork and re-create a new community here.