GSoC 2016 Week 10: Android SDK Tools in Debian

[Chirayu Desai <chirayudesai1@gmail.com>]

Hello,

This week I worked on creating a new source package,
src:android-framework-23 [1], which is used to build "android.jar". It
is like the header file for android apps, containing the declaration
of android's public API, and as such is a crucial component of the
SDK.
[1]: https://gitlab.com/cde/debian_android-tools_android-framework-23

I started by creating the new source package and importing the
multiple upstream tarballs. While that had been done in the past for
other android-tools repositories, I made the Makefile here much more
generic so that it becomes very easy to import a new repo whenever
required [2]. I've been trying to follow that in general while working
on this project, to make things more maintainable in the future.
[2]: https://gitlab.com/cde/debian_android-tools_android-framework-23/blob/master/debian/rules#L6

Another student, Kai-Chung Yan, had already done some work on this
before, so I started with importing that (it was previously done in
another source package, src:android-platform-frameworks-base, however
android.jar is built using source code from multiple upstream git
repositories, so we decided to create a new source package). And then
I made changes to get it to build in the new source package, and
included source code from the other upstream tarballs I imported.

I haven't filed an ITP for the new package yet, as there are still
some improvements I would like to make the upstream update process
more automated and less manual for this repository. We're hitting a
bug in mk-origtargz [3] which results in a error when we try to
exclude files from the upstream tarball, and hence for now we need to
import it manually which is error prone.
[3]: https://bugs.debian.org/831870

Having a working "android.jar" brings us closer to the goal of
building an android app using the SDK packages in Debian, and I'm
confident about getting it done in the next two weeks.

Note: The number of my weekly report (GSoC 2016 Week #N) appears to be
one higher than some other reports, it looks like this is due to me
counting the mid-term evaluation week too (i.e. not skipping any week
since GSoC started).

Regards,
Chirayu Desai