Profile cover photo
Profile photo
Karim Yaghmour
1,144 followers -
On my way to ...
On my way to ...

1,144 followers
About
Karim's interests
View all
Karim's posts

Post has shared content
Tada
Here’s an early look at the next version of Android for testing, development, and feedback. Get your apps ready for Android O: https://goo.gl/hf2I8I

Watch out for the SDK coming later today! 

Post has attachment
My good friend and editor +Andy Oram took the time earlier this year to interview a couple of key people to summarize the history and state of the relationship between the Android team and the Linux kernel community. This is partly what I've been telling people in-person for the past few years. It's nice to see it in writing with input from different points of view.

Post has shared content
Le-sigh. The main issue here seems to be the ability to control the bootloader over the UART (allowing factory reset, and other low-level tools from the bootloader).

Having the FIQ enabled is tough as its a useful debug tool (seemingly designed so that limited but useful debug data could be collected by a store tech when a consumer's phone was acting up). But these days I guess the information leakage is probalby just too risky (though what you can do with that info from only the serial line, I don't know).

Most of the article seems to be covering cases with debug/eng builds (gasp! root shell!) which confuses the degree of severity here for consumer builds.

Its frustrating because I've been wanting the headphone UART or some sort of UART access to be standardized so that folks can debug and hack on their own unlocked devices without needing to break the case open and solder, and this sort of report makes vendors less likely to enable such functionality.

But I guess a counter argument would be if this sort of UART interface was standardized, the security practices around it would also be more standardized, making it less likely that custom vendor bootloader commands would open such a hole.

Post has attachment
Current market watches running Apple Watch OS, Android Wear, Tizen, etc. but can we have Mozilla do one too, and Ubuntu for watches as well?

Post has shared content
Android Platform Tools 25.0.4 released

Android Platform Tools 25.0.4 has been released. You can download it
from https://developer.android.com/studio/releases/platform-tools.html#download
now.

The release notes should be up on that page tomorrow, but in the meantime:

- Added experimental libusb support to linux/mac adb

To use the libusb backend, set the environment variable ADB_LIBUSB=1
before launching adb server. The new `adb host-features` command will
tell you whether or not you're using libusb. In this release, the old
non-libusb implementation remains the default.

- fastboot doesn't hang 2016 MacBook Pros anymore

This fixes https://code.google.com/p/android/issues/detail?id=231129.

As usual, file bugs at b.android.com.



But... the really good news for those who complained before about the lack of notifications is that there's now a android-platform-tools-announce@ group. Here's this same post there as proof: https://groups.google.com/forum/#!topic/android-platform-tools-announce/VnKYHlqQCIg

Post has attachment
Fine accommodations in Nuremberg.
PhotoPhotoPhotoPhotoPhoto
3/16/17
5 Photos - View album

Post has attachment
Klingon ship approaching

Post has attachment
My "Embedded Android Workshop with Nougat" at Embedded World 2017.

Post has attachment
My "Android Things: Android for IoT" presentation at Embedded World 2017

Post has attachment
Embedded World has officially been trolled.

CC +Marek Vašut​, +chris simmonds​, +Robert Schwebel​, +Drew Fustini​
Photo
Wait while more posts are being loaded