Canonical Voices

Posts tagged with 'click'

Nicholas Skaggs

Creating multi-arch click packages

Click packages are one of the pieces of new technology that drives the next version of ubuntu on the phone and desktop. In a nutshell click packages allow for application developers to easily package and deliver application updates independent of the distribution release or archive. Without going into the interesting technical merits and de-merits of click packages, this means the consumer can get faster application updates. But much of the discussion and usage of click packages until now has revolved around mobile. I wanted to talk about using click packages on the desktop and packaging clicks for multiple architectures.

The manifest file
Click packages follow a specific format. Click packages contain a payload of an application's libraries, code, artwork and resources, along with its needed external dependencies. The description of the package is found in the manifest file, which is what I'd like to talk about. The file must contain a few keys, but one of the recognized optional keys is architecture. This key allows specifying architectures the package will run on.

If an application contains no compiled code, simply use 'all' as the value for architecture. This accomplishes the goal of running on all supported architectures and many of the applications currently in the ubuntu touch store fall into this category. However, an increasing number of applications do contain compiled code. Here's how to enable support across architectures for projects with compiled code.

Fat packages
The click format along with the ubuntu touch store fully support specifying one or more values for specific architecture support inside the application manifest file. Those values follow the same format as dpkg architecture names. Now in theory if a project containing compiled code lists the architectures to support, click build should be able to build one package for all. However, for now this process requires a little manual intervention. So lets talk about building a fat (or big boned!) package that contains support for multiple architectures inside a single click package.

Those who just want to skip ahead can check out the example package I put together using clock. This same package can be found in the store as multi-arch clock test. Feel free to install the click package on the desktop, the i386 emulator and an armhf device.

Building a click for a different architecture
To make a multi-arch package a click package needs to be built for each desired architecture. Follow this tutorial on for more information on how to create a click target for each architecture. Once all the targets are setup, use the ubuntu sdk to build a click for each target. The end result is a click file specific to each architecture.

For example in creating the clock package above, I built a click for amd64, i386 and armhf. Three files were generated:

Notice the handy naming scheme allows for easy differentiation as to which click belongs to which architecture. Next, extract the compiled code from each click package. This can be accomplished by utilizing dpkg. For example,

dpkg -x amd64

Do this for each package. The result should be a folder corresponding to each package architecture.

Next copy one version of the package for use as the base of multi-arch click package. In addition, remove all the compiled code under the lib folder. This folder will be populated with the extracted compiled code from the architecture specific click packages.

cp amd64 multi
rm -rf multi/lib/*

Now there is a folder for each click package, and a new folder named multi that contains the application, minus any compiled code.

Creating the multi-arch click
Inside the extracted click packages is a lib folder. The compiled modules should be arranged inside, potentially inside an architecture subfolder (depending on how the package is built).

Copy all of the compiled modules into a new folder inside the lib folder of the multi directory. The folder name should correspond to the architecture of the complied code. Here's a list of the architectures for ARM, i386, and amd64 respectively.


You can check the naming from an intended device by looking in the application-click.conf file.

grep ARCH /usr/share/upstart/sessions/application-click.conf

To use the clock package as an example again, here's a quick look at the folder structure:


The contents of lib/* from each click package I built earlier is under a corresponding folder inside the multi/lib directory. So for example, the lib folder from became lib/i386-linux-gnu/.

Presto, magic package time! 
Finally the manifest.json file needs to be updated to reflect support for the desired architectures. Inside the manifest.json file under the multi directory, edit the architecture key values to list all supported architectures for the new package. For example to list support for ARM and x86 architectures,

"architecture": ["armhf", "i386", "amd64"],

To build the new package, execute click build multi. The resulting click should build and be named with a prefix. This click can be installed on any of the specified architectures and is ready to be uploaded to the store.

Caveats, nibbly bits and bugs
So apart from click not automagically building these packages, there is one other bug as of this writing. The resulting multi-arch click will fail the automated store review and instead enter manual review. To workaround this request a manual review. Upon approval, the application will enter the store as usual.

In summary to create a multi-arch click package build a click for each supported architecture. Then pull the compiled library code from each click and place into a single click package. Next modify the click manifest file to state all of the architectures supported. Finally, rebuild the click package!

I trust this explanation and example provides encouragement to include support for x86 platforms when creating and uploading a click package to the store. Undoubtedly there are other ways to build a multi-arch click; simply ensure all the compiled code for each architecture is included inside the click package. Feel free to experiment!

If you have any questions as usual feel free to contact me. I look forward to seeing more applications in the store from my unity8 desktop!

Read more

Yesterday’s autopkgtest 3.2 release brings several changes and improvements that developers should be aware of.

Cleanup of CLI options, and config files

Previous adt-run versions had rather complex, confusing, and rarely (if ever?) used options for filtering binaries and building sources without testing them. All of those (--instantiate, --sources-tests, --sources-no-tests, --built-binaries-filter, --binaries-forbuilds, and --binaries-fortests) now went away. Now there is only -B/--no-built-binaries left, which disables building/using binaries for the subsequent unbuilt tree or dsc arguments (by default they get built and their binaries used for tests), and I added its opposite --built-binaries for completeness (although you most probably never need this).

The --help output now is a lot easier to read, both due to above cleanup, and also because it now shows several paragraphs for each group of related options, and sorts them in descending importance. The manpage got updated accordingly.

Another new feature is that you can now put arbitrary parts of the command line into a file (thanks to porting to Python’s argparse), with one option/argument per line. So you could e. g. create config files for options and runners which you use often:

$ cat adt_sid

$ adt-run libpng @adt_sid

Shell command tests

If your test only contains a shell command or two, or you want to re-use an existing upstream test executable and just need to wrap it with some command like dbus-launch or env, you can use the new Test-Command: field instead of Tests: to specify the shell command directly:

Test-Command: xvfb-run -a src/tests/run
Depends: @, xvfb, [...]

This avoids having to write lots of tiny wrappers in debian/tests/. This was already possible for click manifests, this release now also brings this for deb packages.

Click improvements

It is now very easy to define an autopilot test with extra package dependencies or restrictions, without having to specify the full command, using the new autopilot_module test definition. See /usr/share/doc/autopkgtest/ for details.

If your test fails and you just want to run your test with additional dependencies or changed restrictions, you can now avoid having to rebuild the .click by pointing --override-control (which previously only worked for deb packages) to the locally modified manifest. You can also (ab)use this to e. g. add the autopilot -v option to autopilot_module.

Unpacking of test dependencies was made more efficient by not downloading Python 2 module packages (which cannot be handled in “unpack into temp dir” mode anyway).

Finally, I made the adb setup script more robust and also faster.

As usual, every change in control formats, CLI etc. have been documented in the manpages and the various READMEs. Enjoy!

Read more

We currently use completely different methods and tools of building test beds and running tests for Debian vs. Click packages, for normal uploads vs. CI airline landings vs. upstream project merge proposal testing, and keep lots of knowledge about Click package test metadata external and not easily accessible/discoverable.

Today I released autopkgtest 3.0 (and 3.0.1 with a few minor updates) which is a major milestone in unifying how we run package tests both locally and in production CI. The goals of this are:

  • Keep all test metadata, such as test dependencies, commands to run the test etc., in the project/package source itself instead of external. We have had that for a long time for Debian packages with DEP-8 and debian/tests/control, but not yet for Ubuntu’s Click packages.
  • Use the same tools for Debian and Click packages to simplify what developers have to know about and to reduce the amount of test infrastructure code to maintain.
  • Use the exact same testbeds and test runners in production CI than what developers use locally, so that you can reproduce and investigate failures.
  • Re-use the existing autopkgtest capabilities for using various kinds of testbeds, and conversely, making all new testbed types immediately available to all package formats.
  • Stop putting tests into the Ubuntu archive as packages (such as mediaplayer-app-autopilot). This just adds packaging and archive space overhead and also makes updating tests a lot harder and taking longer than it should.

So, let’s dive into the new features!

New runner: adt-virt-ssh

We want to run tests on real hardware such as a laptop of a particular brand with a particular graphics card, or an Ubuntu phone. We also want to restructure our current CI machinery to run tests on a real OpenStack cloud and gradually get rid of our hand-maintained QA lab with its test machines. While these use cases seem rather different, they both have in common that there is an already existing machine which is pretty much only accessible with ssh. Once you have an ssh connection, they look pretty much the same, you just need different initial setup (like fiddling with adb, calling nova boot, etc.) to prepare them.

So the new adt-virt-ssh runner factorizes all the common bits such as communicating with adt-run, auto-detecting sudo availability, doing SSH connection sharing etc., and delegates the target specific bits to a “setup script”. E. g. we could specify --setup-script ssh-setup-nova or --setup-script ssh-setup-adb which would then get called with open at the appropriate time by adt-run; it calls the nova commands to create a VM, or run a few adb commands to install/start ssh and install the public key. Then autopkgtest does its thing, and eventually calls the script with cleanup again. The actual protocol is a bit more involved (see manpage), but that’s the general idea.

autopkgtest now ships readymade scripts for these two use cases. So you could e. g. run the libpng tests in a temporary cloud VM:

# if you don't have one, create it with "nova keypair-create"
$ nova keypair-list
| pitti | 9f:31:cf:78:50:4f:42:04:7a:87:d7:2a:75:5e:46:56 |

# find a suitable image
$ nova image-list 
| ca2e362c-62c9-4c0d-82a6-5d6a37fcb251 | Ubuntu Server 14.04 LTS (amd64 20140607.1) - Partner Image                         | ACTIVE |  

$ nova flavor-list 
| 100 | standard.xsmall  | 1024      | 10   | 10        |      | 1     | 1.0         | N/A       |

# now run the tests: please be patient, this takes a few mins!
$ adt-run libpng --setup-commands="apt-get update" --- ssh -s /usr/share/autopkgtest/ssh-setup/nova -- \
   -f standard.xsmall -i ca2e362c-62c9-4c0d-82a6-5d6a37fcb251 -k pitti
adt-run [16:23:16]: test build:  - - - - - - - - - - results - - - - - - - - - -
build                PASS
adt-run: @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ tests done.

Please see man adt-virt-ssh for details how to use it and how to write setup scripts. There is also a commented /usr/share/autopkgtest/ssh-setup/SKELETON template for writing your own for your use cases. You can also not use any setup script and just specify user and host name as options, but please remember that the ssh runner cannot clean up after itself, so never use this on important machines which you can’t reset/reinstall!

Test dependency installation without apt/root

Ubuntu phones with system images have a read-only file system where you can’t install test dependencies with apt. A similar case is using the “null” runner without root. When apt-get install is not available, autopkgtest now has a reduced fallback mode: it downloads the required test dependencies, unpacks them into a temporary directory, and runs the tests with $PATH, $PYTHONPATH, $GI_TYPELIB_PATH, etc. pointing to the unpacked temp dir. Of course this only works for packages which are relocatable in that way, i. e. libraries, Python modules, or command line tools; it will totally fail for things which look for config files, plugins etc. in hardcoded directory paths. But it’s good enough for the purposes of Click package testing such as installing autopilot, libautopilot-qt etc.

Click package support

autopkgtest now recognizes click source directories and *.click package arguments, and introduces a new test metadata specification syntax in a click package manifest. This is similar in spirit and capabilities to DEP-8 debian/tests/control, except that it’s using JSON:

    "x-test": {
        "unit": "tests/unittests",
        "smoke": {
            "path": "tests/smoketest",
            "depends": ["shunit2", "moreutils"],
            "restrictions": ["allow-stderr"]
        "another": {
            "command": "echo hello > /tmp/world.txt"

For convenience, there is also some magic to make running autopilot tests particularly simple. E. g. our existing click packages usually specify something like

    "x-test": {
        "autopilot": "ubuntu_calculator_app"

which is enough to “do what I mean”, i. e. implicitly add the autopilot test depends and run autopilot with the specified test module name. You can specify your own dependencies and/or commands, and restrictions etc., of course.

So with this, and the previous support for non-apt test dependencies and the ssh runner, we can put all this together to run the tests for e. g. the Ubuntu calculator app on the phone:

$ bzr branch lp:ubuntu-calculator-app
# built straight from that branch; TODO: where is the official" download URL?
$ wget
$ adt-run ubuntu-calculator-app/ --- \
      ssh -s /usr/share/autopkgtest/ssh-setup/adb
Traceback (most recent call last):
  File "/tmp/adt-run.KfY5bG/tree/tests/autopilot/ubuntu_calculator_app/tests/", line 93, in test_divide_with_infinity_length_result_number
  File "/tmp/adt-run.KfY5bG/tree/tests/autopilot/ubuntu_calculator_app/tests/", line 63, in _assert_result
    self.main_view.get_result, Eventually(Equals(expected_result)))
  File "/usr/lib/python3/dist-packages/testtools/", line 406, in assertThat
    raise mismatch_error
testtools.matchers._impl.MismatchError: After 10.0 seconds test failed: '0.33333333' != '0.3'

Ran 33 tests in 295.586s
FAILED (failures=1)

Note that the current adb ssh setup script deals with some things like applying the autopilot click AppArmor hooks and disabling screen dimming, but it does not do the first-time setup (connecting to network, doing the gesture intro) and unlocking the screen. These are still on the TODO list, but I need to find out how to do these properly. Help appreciated!

Click app tests in schroot/containers

But, that’s not the only thing you can do! autopkgtest has all these other runners, so why not try and run them in a schroot or container? To emulate the environment of an Ubuntu Touch session I wrote a --setup-commands script:

adt-run --setup-commands /usr/share/autopkgtest/setup-commands/ubuntu-touch-session \
    ubuntu-calculator-app/ --- schroot utopic

This will actually work in the sense of running (and succeeding) the autopilot tests, but it will fail due to a lot of libust[11345/11358]: Error: Error opening shm /lttng-ust-wait... warnings on stderr. I don’t know what these mean, just that I also see them on the phone itself occasionally.

I also wrote another setup-commands script which emulates “read-only apt”, so that you can test the “unpack only” fallback. So you could prepare a container with click and the App framework preinstalled (so that it doesn’t always take ages to install them), starting from a standard adt-build-lxc container:

$ sudo lxc-clone -o adt-utopic -n click
$ sudo lxc-start -n click
  # run "sudo apt-get install click ubuntu-sdk-libs ubuntu-app-launch-tools" there
  # then "sudo powerdown"

# current apparmor profile doesn't allow remounting something read-only
$ echo "lxc.aa_profile = unconfined" | sudo tee -a /var/lib/lxc/click/config

Now that container has enough stuff preinstalled to be reasonably fast to set up, and the remaining test dependencies (mostly autopilot) work fine with the unpack/$*_PATH fallback:

$ adt-run --setup-commands /usr/share/autopkgtest/setup-commands/ubuntu-touch-session \
          --setup-commands /usr/share/autopkgtest/setup-commands/ro-apt \
          ubuntu-calculator-app/ \
          --- lxc -es click

This will successfully run all the tests, and provided you have apt-cacher-ng installed, it only takes a few seconds to set up. This might be a nice thing to do on merge proposals, if you don’t have an actual phone at hand, or don’t want to clutter it up.

autopkgtest 3.0.1 will be available in Utopic tomorrow (through autosyncs). If you can’t wait to try it out, download it from my people.c.c page ☺.

Feedback appreciated!

Read more
Nicholas Skaggs

Building click packages should be easy. And to a reasonable extent, qtcreator and click-buddy do make it easy. Things however can get a bit more complicated when you need to build a package that needs to run on an armhf device (you know like your phone!). Since your pc is almost certainly based on x86, you need to use, create or fake an armhf environment for building the package.

So then what options exist for getting a proper build of a project that will install properly on your device?

A phone can be more than a phone
It can also be a development environment!? Although it's not my recommendation, you can always use the source device to compile the package with. The downsides of this is namely speed and storage space. Nevertheless, it will build a click.

  1. shell into your device (adb shell / ssh mydevice)
  2. checkout the code (bzr branch lp:my-project)
  3. install the needed dependencies and sdk (apt-get install ubuntu-sdk)
  4. build with click-buddy (click-buddy --dir .)
Chroot to the rescue
The click tools contain a handy way to build a chroot expressly suited for use with click-buddy to build things. Basically, we can create a nice fake environment and pretend it's armhf, even though we're not running that architecture.

sudo click chroot -a armhf -f ubuntu-sdk-14.04 create
click-buddy --dir . --arch armhf

Most likely your package will require extra dependencies, which for now will need to be specified and passed in with the --extra-deps argument. These arguments are packages names, just like you would apt-get. Like so;

click-buddy --dir . --arch armhf --extra-deps "libboost-dev:armhf libssl-dev:armhf"

Notice we specified the arch as well, armhf. If we also add a --maint-mode, our extra installed packages will persist. This is handy if you will only ever be building a single project and don't want to constantly update the base chroot with your build dependencies.

Qtcreator build it for me!
Cmake makes all things possible. Qt Creator can not only build the click for you, it can also hold your hand through creating a chroot1. To create a chroot in qtcreator, do the following:
  1. Open Qt Creator
  2. Navigate to Tools > Options > Ubuntu > Click
  3. Click on Create Click Target
  4. After the click target is finished, add the dependencies needed for building. You can do this by clicking the maintain button.  
  5. Apt-get add what you need or otherwise setup the environment. Once ready, exit the chroot.
Now you can use this chroot for your project
  1. Open qt creator and open the project
  2. Select armhf when prompted
    1. You can also manually add the chroot to the project via Projects > Add kit and then select the UbuntuSDK armhf kit.
  3. Navigate to Projects tab and ensure the UbuntuSDK for armhf kit is selected.
  4. Build!
Rolling your own chroot
So, click can setup a chroot for you, and qt creator can build and manage one too. And these are great options for building one project. However if you find yourself building a plethora of packages or you simply want more control, I recommend setting up and using your own chroot to build. For my own use, I've picked pbuilder, but you can setup the chroot using other tools (like schroot which Qt Creator uses).

sudo apt-get install qemu-user-static ubuntu-dev-tools
pbuilder-dist trusty armhf create
pbuilder-dist trusty armhf login --save-after-login

Then, from inside the chroot shell, install a couple things you will always want available; namely the build tools and bzr/git/etc for grabbing the source you need. Be careful here and don't install too much. We want to maintain an otherwise pristine environment for building our packages. By default changes you make inside the chroot will be wiped. That means those package specific dependencies we'll install each time to build something won't persist.

apt-get install ubuntu-sdk bzr git phablet-tools

By exiting, you'll notice pbuilder will update the base tarball with our changes. Now, when you want to build something, simply do the following:

pbuilder-dist trusty armhf login
bzr branch lp:my-project
apt-get install build-dependencies-you-need

Now, you can build as usual using click tools, so something like

click-buddy --dir .

works as expected. You can even add the --provision to send the resulting click to your device. If you want to grab the resulting click, you'll need to copy it before exiting the chroot, which is mounted on your filesystem under /var/cache/pbuilder/build/. Look for the last line after you issue your login command (pbuilder-dist trusty armhf login). You should see something like, 

File extracted to: /var/cache/pbuilder/build//26213

If you cd to this directory on your local machine, you'll see the environment chroot filesystem. Navigate to your source directory and grab a copy of the resulting click. Copy it to a safe place (somewhere outside of the chroot) before exiting the chroot or you will lose your build! 

But wait, there's more!
Since you have access to the chroot while it's open (and you can login several times if you wish to create several sessions from the base tarball), you can iteratively build packages as needed, hack on code, etc. The chroot is your playground.

Remember, click is your friend. Happy hacking!

1. Thanks to David Planella for this info

Read more

Here's how I build and test click packages on the target devices, note that this only works for packages that don't need to be compiled from their own sources:

# e.g.; branch=lp:dropping-letters
# e.g.; click_package=com.ubuntu.dropping-letters
# e.g.; test=dropping_letters

bzr branch lp:~sergiusens/+junk/click_ready $tool_path

bzr branch $branch branch
cd branch
$tool_path/ --bzr-source $branch

adb push *.click /tmp
adb shell 'sudo -u phablet pkcon install-local *.click'
# this may require lp:phablet-tools
phablet-tools autopilot --dbus-probe enable

phablet-click-test-setup --click $click_package

phablet-test-run $test

I hope to make this a bzr plugin to avoid all these instructions. For certain we need to have a session at the upcoming session to discuss better testing strategies that would satisfy testing:

  • On production images.
  • Staying in readonly.
  • Adding test dependencies.
  • Best way to bundle tests for click.
  • Best ways to provide testing tools not included in the images.

The tools suffice for the moment but this can be made more elegant.

Read more

In the recent images, starting the 20130813.1, you will notice that click packages have started to show up on the built images. That is, from a user perspective you would probably not notice at all, but if you dive into the system, you will.

What and how it is built

Currently, all the packages for the Core Community Apps are being built as click packages, as these are the easiest since are binary independent.

The click packages are built on jenkins, you can take a look at the sudoku-app. These are built in some dynamic form with a tool that basically reads out the current debian packaging from the sources and builds out a manifest and puts everything in a staged location on which we just run click build.

Since there is no concept of a PPA for a click package they are provisioned on one of the Canonical Servers that can be found here.

What's installed

While all of the click packages could be installed on the image, we are only installing 3:

  • dropping letters.
  • sudoku.
  • stock ticker.

The installation is done from a hook in livecd-rootfs which installs them as the phablet user in order for you to enjoy them on the install.

The list of packages to install is today sort of hardcoded in click_list which is built out by Any new addtions, to this date will come from here.

What happens when a click package is installed

When you install a click package you need to target a user, a click package with no hooks would just lay it out plainly. A click package manifest with hooks will trigger certain actions upon installation, today there are two important ones:

  • security hook: this creates the apparmor profile for the app to run under confinement, these land in /var/lib/apparmor/profiles/.
  • desktop hook: this is a user hook, which basically provisions a desktop file modified for confinement and with the actual path to where the application lives, this lands in ~/.local/share/applications. The target is to have unity8 not use the desktop files but to eventually launch them through upstart.

To see the click packages installed for your user (phablet), just run click list. If you want to see the manifests for these click packages just run click list --manifest

Where to get more info?

There's isn't much lying around, but Colin has made info available on readthedocs, with regards to security, a good starting point is the security team's wiki.

Read more