Why don't they just target 3.9?
Actually I was hoping they would drop their power management patches and they have! If one looks in /kernel/power the only file not in linus's tree is suspend_time.c and that's encouraging. Hopefully this code has been tested and is as power efficient as Android's current code if not more so.
PS: It will be interesting to see what google does with F2FS too.
Is there an easy overview somewhere with the diff sizes between the Android kernel and Linus' branch with the corresponding version number and how this has evolved over the kernel releases?
Also, it would be interesting to see those diffs divided into the different directories/functional units of the kernel sources.
In a way that would give an indication of how long they have to go to run on an almost pure mainline kernel.
Actually I like the changes in Java 7.
* Can use strings in cases in switch()
* Type interfering, less code, less repetition of code
Unified ARM sounds really nice. Hopefully that makes it easier for CyanogenMod, Raspberry Pi, Debian/Ubuntu and everyone.
Ah yes... It is sooooo difficult to just take vanilla linux, all the drivers, systemd, dbus, pulseaudio, port current homescreen design to Enlightenment (ship that shit?), slowly phase out Java (like Microsoft phases out win32), copy Tizen work on having HTMLv5 work on EFL, rename Google Play to Repository browser and put Google Play inside a repository file... And just fscking ship it with GNU tools.
DRM the shit out of Coreboot and have Android as a secure payload. Allow other unsigned payloads, but warn about security (and have a checkbox saying "Do not warn me about this payload anymore"). Everybody happy.
Google sucks. So much so that I've ditched even search for DuckDuckGo.