First Signs Of Wayland Running On Android

[daniels]

It makes it possible to test wayland-apps on android right now? That's useful if we eventually get a more direct wayland path on android (or plain linux on the same hardware), since it means we'll have slightly better-tested software available. It also makes it possible to run wayland apps on top of android, if you should want that (sort of like why you might want an X server in MacOS X).

I'm not really sure how it's possible to get a more direct than Weston drawing directly to the framebuffer as it does now.

[pq__]

What's the objective of running wayland on something that once ran Android?
I.e., what you end up with here is....
hardware --> kernel --> display driver --> wayland --> ??? --> acid trip --> profit
instead of...
hardware --> kernel --> display driver --> android

Note: It is no longer Android once you replace Android with wayland....

That's actually a fairly accurate description, replacing Android with Wayland, because pretty much every Android system service and interface directly or (most often) indirectly depends on SurfaceFlinger. Even the power button, whose action is to show a dialog for shutdown/something/cancel/whatever, and it cannot show anything without SurfaceFlinger, so it just doesn't work anymore.

The objective is to offer an alternative for Android, and take advantage of the huge amount of devices and drivers already written for them. In the future we can start getting familiar programs on those devices. That is what open source is.

89c51, it has been Collabora's own project.

Like daniels said, there is no more direct path that what I did: Weston runs directly on top of the framebuffer. It is taking SurfaceFlinger's place.

Running existing Android stuff on top of Wayland is a whole another story, and I personally do not have plans for it. It would be a huge project about rewriting several Android components, and what would be even more difficult, interfacing to existing Android system services in proper ways.

[89c51]

89c51, it has been Collabora's own project.

thanks for the answer

[droidhacker]

That sounds about equivalent to you reading "First Signs Of Wayland Running On Linux" and saying "It's no longer Linux once you replace Linux with wayland."

Doesn't make any sense.

He didn't replace Android with Wayland. He replaced Android's UI with Wayland, much like Wayland is being made to work in place of X on Linux. And similarly, I'd guess the plan is to make Android's UI work through Wayland as X works through Wayland. Or maybe modify Android to do all its output directly through Wayland.

If you understood Android, you would understand why you are incorrect. Basically, there is a kernel and a few basic utilities, everything above that is Android. As pq__ stated, because of the dependencies, once you change out Android's "UI", all that's left is Linux and a few utilities. Android is gone.

It is still, however, Linux....

[pq__]

If you understood Android, you would understand why you are incorrect. Basically, there is a kernel and a few basic utilities, everything above that is Android. As pq__ stated, because of the dependencies, once you change out Android's "UI", all that's left is Linux and a few utilities. Android is gone.

It is still, however, Linux....

From a user perspective, Android is gone, yes. But there is still some serious infrastructure left: the build tools and toolchains, installing and debugging tools (adb etc.), HALs, bootloaders(?), and hardware "support" that is not in upstream Linux and might never be. In my opinion, calling that "Linux and a few utilities" is an understatement. But that is a developer's point of view.

[89c51]

would there be any benefits if android replaced its WM with Wayland and ported all thats above to it???

(excuse me if this doesn't make any sense, i am no dev)

[pq__]

would there be any benefits if android replaced its WM with Wayland and ported all thats above to it???

Sorry, I don't really know Android well enough to answer that. An even better question would be: which one would be worth more, migrating to Wayland or cleaning up their existing code base? Someone might argue that both options would require a full rewrite of major parts...

Someone might also consider that migrating to a "standard technology" (Wayland) from the respective Android-exclusive ad-hoc technology parts to be a benefit in itself.