Android-x86 6.0-RC1 Released With OpenGL ES Via Mesa, Linux 4.4.12

[phoronix]

Phoronix: Android-x86 6.0-RC1 Released With OpenGL ES Via Mesa, Linux 4.4.12

After months of Android-x86 6.0 test releases, the Android-x86 "Marshmallow" release is up to the release candidate phase...

http://www.phoronix.com/scan.php?page=news_item&px=Android-x86-6.0-RC1

[starshipeleven]

oh nice that was fast, finally emulating shitty embedded device GPU features on x86 CPU, this should mean we can run this pretty much everywhere.
Downloading now.

[droidhacker]

This will immediately become a lot more interesting when it is Android-N. Until then, Android really isn't ideal for a desktop system since it wastes so much screen space. With N though, and native floating multi-window (which a lot of people don't realize is actually in there... all they know about is split screen), a couple of simple hacks will make it into a usable desktop OS.

Hack 1: DEFAULT into multi-window mode.
Hack 2: application switcher that is built into the nav bar, instead of needing to go through the square button application switcher, which is equivalent to gnome3's crazy corner that makes your head explode. The square button screen is ok for small screen devices, but absolutely useless when it gets to large screen.

[eydee]

So let's see how Candy Crush and Angry Birds run on the GTX 1080...

[starshipeleven]

This will immediately become a lot more interesting when it is Android-N. Until then, Android really isn't ideal for a desktop system since it wastes so much screen space. With N though, and native floating multi-window (which a lot of people don't realize is actually in there... all they know about is split screen), a couple of simple hacks will make it into a usable desktop OS.

I feel an announcement coming. Let me make some space:

THE! YEAR! OF! THE! ANDROID! DESSSSKTOPPPPP!!!!!11!

Hack 1: DEFAULT into multi-window mode.
Hack 2: application switcher that is built into the nav bar, instead of needing to go through the square button application switcher, which is equivalent to gnome3's crazy corner that makes your head explode. The square button screen is ok for small screen devices, but absolutely useless when it gets to large screen.

Currently this is done already by RemixOS, a chinese closed-source Android x86 distro based on this project that enables such experimental features and then adds some of their own stuff.

http://www.jide.com/remixos

Awesome, but I'm NOT trusting them in the slightest.

That's what I want to have on a desktop tho.

[hrkristian]

So does anyone have experience compiling this for existing Android x86 tablets?
I've been wanting to push this thing onto an abandoned ASUS tablet but never had the time nor known to even look to make it happen...

[starshipeleven]

So does anyone have experience compiling this for existing Android x86 tablets?
I've been wanting to push this thing onto an abandoned ASUS tablet but never had the time nor known to even look to make it happen...

Existing x86 Android tablets are more similar to ARM boards from a ROM maker point of view, as they lack a BIOS and all.

Custom ROMs for x86 stuff are rare as afaik there is a costly Intel SDK or something to get first to develop for that target at all (as it's not a matter of architecture but of figuring out how to initialize that specific board hardware and getting the blobs and other stuff).

Best bet is looking on XDAdevelopers if someone made custom ROMs for that model, but in general it's MUCH easier to get this in a Windows tablet than in an Android one.

[starshipeleven]

Hell yeah.
Installed on a usb 3.0 thumbdrive, my touchscreen netbook likes it a lot.
Runs smoothly for documents and web and media, it somehow ships with google play pre-installed.

Will try out some games and see how it goes.

Has anyone tried it in Virtualbox/KVM/whatever?

If it keeps going like this I'm going to buy a windows tablet to install this in.

[droidhacker]

Currently this is done already by RemixOS, a chinese closed-source Android x86 distro based on this project that enables such experimental features and then adds some of their own stuff.

http://www.jide.com/remixos

Awesome, but I'm NOT trusting them in the slightest.

That's what I want to have on a desktop tho.

The purpose of remix is to provide control over the world's desktops to the chinese government by adding floating multi-window support, which Android N has natively WITHOUT remix's bag-of-malware.

I'm not talking about adding floating multi-window, I'm talking about managing it. This is a very superficial level of work that I would implement myself in about an hour or less if the N source code was public, which it isn't yet.

[droidhacker]

Existing x86 Android tablets are more similar to ARM boards from a ROM maker point of view, as they lack a BIOS and all.

Custom ROMs for x86 stuff are rare as afaik there is a costly Intel SDK or something to get first to develop for that target at all (as it's not a matter of architecture but of figuring out how to initialize that specific board hardware and getting the blobs and other stuff).

Best bet is looking on XDAdevelopers if someone made custom ROMs for that model, but in general it's MUCH easier to get this in a Windows tablet than in an Android one.

That isn't really a valid comparison. While it may be true that Android x86 adds in a few early startup and configuration components (i.e., partition/format the disk, install grub), the reality is that Android itself doesn't give a rat's ass if its being started by grub or littlekernel or whatever else you come up with. Once the control is handed over to the kernel, it deals with the rest of everything on its own.

Most of what the android x86 project deals with are hardware interfacing related, such as getting surfaceflinger to play nicely with mesa.

Fun fact: 96boards hardware (such as the Dragonboard 410C) run multiple software stacks. In fact, the DB410C has TWO Android variants; the Qualcomm variant, which is a more traditional setup with the "official" binary drivers, and the Linaro variant, which uses a recent kernel and freedreno. The Linaro variant is predominantly upstream for both kernel and Android, and leverages very strongly on the Android x86 project. Yes, this is despite the fact that DB410C is an ARM board.

I think that it is important to recognize that the Android x86 project isn't really all about x86. It is about breaking the Android dependence on binary drivers and opening it up to *ALL* of the hardware that is supported by the upstream kernel, and in the MANNER that that hardware is supported by the kernel. So what we end up with are a bunch of HALs that are a heck of a lot more generic than what you would typically find in Android.