Akademy/2023/Plasma Mobile BoF

From KDE Community Wiki
< Akademy‎ | 2023

Meeting Notes

Plasma Mobile and Qt 6

  • Migration mostly done

Onboarding

  • We need more people on this
  • What makes it hard to contribute to Plasma Mobile?
  • A lot of work can be done on desktop, doing things on hardware is quite hard at the moment, and for some things that is necessary.
  • Would an easy cross-compilation setup help? (e.g. via kdesrc-build)
  • Different use-cases
    • Application development, ideally a single click in the IDE bringing your app on the phone
    • Base image/shell development
  • Cross-compilation complexity vs emulated/native compilation performance
    • binfmt emulation works also in combination with Docker
    • this needs more documentation
  • weekly/nightly postmarketOS images would help a lot
    • postmarketOS not using glibc and systemd can be a challenge

Hardware support

  • Primarily used hardware are the One+ 6, works better than the PinePhone.
    • those devices are aging, makes it harder to be motivated to work on this for your own use
  • Fairphone could be an option
  • mainline Linux situation on random current phones has improved for the basics, camera etc still a challenge

Apps

  • consider Android as a stepping stone/proving ground for Plasma Mobile apps
  • less useful for apps using highly platform-specific APIs (dialer, calendar, KIO, etc)
    • Merkuro is missing due to Akonadi
    • kclock is missing due to maintainer decision
    • Calindori and Plasma Phonebook work, but don't interact with the Android platform contact/calendar
  • Android apps not discoverable (not in Play Store and not in Google Play)
  • F-Droid:
    • client fails when APK changed since the last repo update
    • stable APKs are rebuilt daily, we don't have a proper release process as for other platforms
  • Google
    • build pipeline into Play Store exist, but requires manual release steps, as well as some initial paperwork

PIM

  • is the current stack around Akonadi the right approach?
  • there is a ton of stuff in there we don't want to rewrite
  • Carl is working on optimizing the multi-process aspect
  • Dan is looking into replace MySQL with Sqlite in Akonadi
  • not something we want on Android though, there we should use the platform services
  • we'd also need sliding window sync in Akonadi (locally only having the last N days of state), which would probably also solve some of the performance problems by massively reducing the amount of local data

Power management

  • push notifications/prepare apps for suspending/killing background apps
  • push notifications need the server deployment and the XDG portal API, everything else should ready meanwhile