Thursday, February 14, 2013

Some quality resources

A couple posts ago, I mentioned the ubuntu quality team was looking for people to join the team and help out in the testing efforts we undertake. Thanks to those of you who've already answered the call and our now joining our testing ranks. We love sharing the joys of testing with others!

We're serious about wanting to make sure you are able to contribute and join the community as easily as possible. So for the last couple months, as a team we've been writing tutorials, giving classroom sessions, and hosting testing events. We really do want you as part of the team. Check out some of the resources available to you and consider becoming a part of the team!

Classroom Sessions
Video Tutorials
Written Walkthroughs

Friday, February 8, 2013

Cadence Weeks, Quality, and You!

Starting tomorrow February 9th, 2013 (heh, some of you reading this might already be in tomorrow), the quality community team will start testing for cadence week #6. During this week, we as a team try and help test specific packages looking for regressions, doing new feature or hardware testing, and making sure our images are in good shape. If your still confused, there's a nice wiki page that lays out what "cadence" means in a bit more detail.

So what does this mean for you, dear reader? Well, we as a team would like you to be involved in helping us test! Everyone has unique ways of interacting with software, and naturally no two computer setups are exactly the same between us. Now, I know what your thinking -- how can I help? I'm no tester, and I don't run development versions of ubuntu!

That's ok! You can still help test without needing to compromise your system. If you don't want to install the development version on your machine, you can use a virtual machine installation instead. If you are unable to run virtual machines, or are confused at the idea, you can still help test by simply running a live session and executing tests there. It's not too hard for you! Check out this walk-through for participating using only an image of the development version of ubuntu and your computer.

To help demonstrate how you can participate, I'll be hosting two live events this next week where I'll be on-hand running through the cadence week tests along with others from the quality team. There will even be a livestream, so if your a visual person (like me!), you can see for yourself how you might be able to contribute. Here's the dates you need to know:

Monday Feb 11th, 1800-1900 UTC in #ubuntu-quality. I'll also be streaming live my participation in executing the tests .

Thursday Feb 14th, 1400-1500 UTC in #ubuntu-quality. No stream, but we'll be hanging out answering questions, and working on submitting test results.


Please consider attending a session or watching the video of the stream afterwards. If you can download an image and boot your computer, you can help test. You want to be a part of ubuntu; let us help you contribute!

Tuesday, February 5, 2013

PSA: Ubuntu Quality wants you!


NOTICE: To whom it may concern, the ubuntu quality team is seeking those with a desire to help ubuntu to contribute to the quality and testing efforts. With a little time and a willingness to learn, you too can unlock the tester within you!

Interested? Please inquire below!

If that text didn't get you, I hope the picture did. Seriously though, if you are here reading this page, I want to offer you an opportunity to help out. We as a team have expanded our activities and projects this cycle and we want to extend an offer for you to come along and learn with us. We're exploring automated testing with autopilot and autopkg, manual testing of images, and the virtues of testing in a regular cadence.

But we can't do it alone, nor do we wish to! We'd love to hear from you. Please have a look at our getting involved page (but do excuse the theme dust!) and get in touch. I offered a challenge to this community in the past, and I was blown away by the emails that flooded my inbox. Send me an email, tell me your interests, and ask me how you can help. Let me help get you started. Flood my inbox again1. Let's make ubuntu better, together!

1. If anyone is counting, I believe the record is ~100 emails in one 24 hour period :-p

Thursday, January 24, 2013

Introspecting with Autopilot

If you remember our post from last time, we had gone through writing our first testcase for firefox, converting a simple manual testcase to utilize autopilot instead. In this post I'd like to talk about how introspection can be used to perform some more complicated automated testcases.

First of all, let's briefly define what we mean by introspection. Specifically, we're talking about introspecting the dbus session for an application on our screen. Trust me, it sounds worse than it is. I'll let you do your own googling if you are curious to learn more. For the rest of us, let's just have a look visually at what we're talking about :-)

If you've got autopilot installed (check out the previous post; install autopilot ppa, sudo apt-get install python-autopilot), you should be able to launch the visualization tool.

autopilot vis

A window should launch, and allow you to select a connection.  This allows you to select which application you wish to introspect. Go ahead and select 'Unity'. If the bareness of the tool scares you, remember it's a development aid, not your browser ;-)

Ok, under the Tree node, you should find a giant list of of nodes and properties for Unity. It may come as a surprise that your desktop is providing this much data about what's going on right now. For instance, have a look under PanelController->Indicators. There's entries for each indicator you are running, along with properties about each one. Ok, so what if your not running Unity? Or, for our purposes, you wish to write a test about an application on our desktop?

Never fear, we can use another feature of autopilot to help launch and introspect an application using the same tool. Go ahead and close the visualization window and enter the following.

autopilot launch gedit
autopilot vis

Notice now we have a new connection called 'Root'. Select it and you'll see the node tree for the gedit window you just launched. The amount of nodes spawned is a bit overwhelming, but you can now use this data to make assertions about what's going on when you interact with the application.

As a quick example, let's say I wanted to know the size of the current gedit window. I we look under 'GeditWindow->globalRect' we can see the current position, and infer the size of the window as well. We can also see things like the title, is_active, and other 'xwindow typish' properties.

In addition, I can find out what buttons are present on the gedit toolbar of the current gedit window. I we look under 'GeditWindow->GtkToolbar' we can see several GtkToolButton nodes. Each has a set of properties, including a name and label. 

So, let's put it all together for a quick example.
 
bzr branch lp:ubuntu-autopilot-tests

Inside the resulting directory you'll notice a geditintrospection folder.

cd ubuntu_autopilot_tests
autopilot run geditintrospection

A gedit window should spawn and disappear -- and hopefully the one testcase should pass. Open up the file geditintrospection/test_geditintrospection.py. Inside you'll notice we're using some of the properties we found while introspecting to show off how we can utilize them to test gedit. Let's cover some of the new functions briefly. You can use the autopilot documentation for more information on what you see.

     def select_single(self, type_name='*', **kwargs):
        """Get a single node from the introspection tree, with type equal to
        *type_name* and (optionally) matching the keyword filters present in
        *kwargs*.


    def select_many(self, type_name='*', **kwargs):
        """Get a list of nodes from the introspection tree, with type equal to
        *type_name* and (optionally) matching the keyword filters present in
        *kwargs*.

  
These two functions allow us to get back the nodes that match our query. For example, you can see the first step of the testcase is to check for a New File button, which is on the gedit toolbar. After asserting it exists, we then click it. Load up gedit for yourself and use the vis tool to confirm. You'll find the node under GeditWindow->GtkBox->GtkToolbar->GtkToolButton. Each button is represented, in this case we pulled the one with label=_New, representing the new file button.

Later we actually interact with gedit by turning on overwrite mode. Normally we would be unable to verify if this indeed worked or not, but you'll notice we once again check for the GtkLabel change from INS to OVR. Again, you can see this under GeditWindow->GtkBox->GeditStatusbar->GtkLabel.

Finally, you'll notice some slight differences from our non-introspection testcase. We now use GtkIntrospectionTestMixin, and launch our application via the launch_test_application function, rather than using AutopilotTestCase. I've shown gtk based examples here, but introspection works with Qt too (using QtIntrospectionTestMixin), so don't be afraid to try it out on any application you are interested in. 

You may also have noticed you branched from a project; ubuntu-autopilot-tests. I'm happy to announce this is the master repository for all the autopilot testcases we'll be writing as a community. Interested in helping contribute? Come join us and get involved! We're tracking work items, including proposed tests for you to work on. In addition, the tests themselves will soon be running on jenkins for everyone in the community to benefit.

Now introspection is still new, and we as a Quality Community team are excited about adopting and utilizing the tool.  There might be some bugs and feature requests (wink, wink autopilot team!) to work out, but we are excited to build a repository of automated tests together.

NOTE: Due to an error during build, it appears the autopilot online documentation is absent or missing pieces. If this occurs, please use the local documentation installed on your machine as part of the autopilot package. You'll find a copy you can browse at /usr/share/doc/python-autopilot/html/index.html.

Thursday, December 13, 2012

Jamming Thursday's!

Right now as I type we have two jams going on! Last week Jono posted about enhancing the ubuntu.com/community page. If your a part of the community, join in raising the banner for your specific focus area. The fun is happening now on #ubuntu-docs. For the full details, see Jono's post. For us quality folks, the pad is here: http://pad.ubuntu.com/communitywebsite-contribute-quality. Feel free to type and edit away!

In addition, as Daniel Holbach mentioned, there is a hackathon for automated testing. Come hang out with us on #ubuntu-quality, learn, ask and write some tests. Again, the full details can be found on Daniel's post.

Come join us!

Wednesday, November 28, 2012

Our first Autopilot testcase

So last time we learned some basics for autopilot testcases. We're going to use the same code branch we pulled now to cover writing an actual testcase.

bzr branch lp:~nskaggs/+junk/autopilot-walkthrough

As a practical example, I'm going to convert our (rather simple and sparse) firefox manual testsuite into an automated test using autopilot. Here's a link to the testcase in question.

If you take a look at the included firefox/test_firefox.py file you should recognize it's basic layout. We have a setup step that launches firefox before each test, and then there are the 3 testcases corresponding to each of the manual tests. The file is commented, so please do have a look through it. We utilize everything we learned last time to emulate the keyboard and mouse to perform the steps mentioned in the manual testcases. Enough code reading for a moment, let's run this thing.

autopilot run firefox

Ok, so hopefully you had firefox launch and run through all the testcases -- and they all, fingers-crossed, passed. So, how did we do it? Let's step through the code and talk about some of the challenges faced in doing this conversion.

Since we want to test firefox in each testcase, our setUp method is simple. Launch firefox and set the focus to the application. Each testcase then starts with that assumption. Inside test_browse_planet_ubuntu we simply attempt to load a webpage. Our assertion for this is to check that the application title changes to "Planet Ubuntu" - - in other words that the page loaded. The other two testcases expand upon this idea by searching wikipedia and checking for search suggestions.

The test_search_wikipedia method uses the keyboard shortcut to open the searchbar, select wikipedia and then search for linux. Again, our only assertion for success here is that the page with a title of Linux and wikipedia loaded. We are unable to confirm for instance, that we properly selected wikipedia as the search engine (although the final assertion would likely fail if this was not the case).

Finally, the test_google_search_suggestions method is attempting to test that the "search suggestions" feature of firefox is performing properly. You'll notice that we are missing the assertion for checking for search suggestions while searching. With the knowledge we're gained up till now, we don't have a way of knowing if the list is generated or not. In actuality, this test cannot be completed as the primary assertion cannot be verified without some way of "seeing" what's happening on the screen.

In my next post, I'll talk about what we can do to overcome the limitations we faced in doing this conversion by using "introspection". In a nutshell by using introspection, autopilot will allow us to "see" what's happening on the screen by interacting with the applications data. It's a much more robust way of "seeing" what we see as a user, rather than reading individual screen pixels. With any luck, we'll be able to finish our conversion and look at accomplishing bigger tasks and tackling larger manual testsuites.

I trust you were able to follow along and run the final example. Until the next blog post, might I also recommend having a look through the documentation and try writing and converting some tests of your own -- or simply extend and play around with what you pulled from the example branch. Do let me know about your success or failure. Happy Testing!

Monday, November 26, 2012

Getting started with Autopilot

If you caught the last post, you'll have some background on autopilot and what it can do. Start there if you haven't already read the post.

So, now that we've seen what autopilot can do, let's dig in to making this work for our testing efforts. A fair warning, there is some python code ahead, but I would encourage even the non-programmers among you to have a glance at what is below. It's not exotic programming (after all, I did it!). Before we start, let's make sure you have autopilot itself installed. Note, you'll need to get the version from this ppa in order for things to work properly:

sudo add-apt-repository ppa:autopilot/ppa
sudo apt-get update && sudo apt-get install python-autopilot

Ok, so first things first. Let's create a basic shell that we can use for any testcase that we want to write. To make things a bit easier, there's a lovely bazaar branch you can pull from that has everything you need to follow along.

bzr branch lp:~nskaggs/+junk/autopilot-walkthrough
cd autopilot-walkthrough

You'll find two folders. Let's start with the helloworld folder. We're going to verify autopilot can see the testcases, and then run and look at the 'helloworld' tests first. (Note, in order for autopilot to see the testcases, you need to be in the root directory, not inside the helloworld directory)

$ autopilot list helloworld
Loading tests from: /home/nskaggs/projects/

    helloworld.test_example.ExampleFunctions.test_keyboard
    helloworld.test_example.ExampleFunctions.test_mouse
    helloworld.test_hello.HelloWorld.test_type_hello_world

 3 total tests.


Go ahead and execute the first helloworld test.

autopilot run helloworld.test_hello.HelloWorld.test_type_hello_world
 
A gedit window will spawn, and type hello world to you ;-) Go ahead and close the window afterwards. So, let's take a look at this basic testcase and talk about how it works.

from autopilot.testcase import AutopilotTestCase

class HelloWorld(AutopilotTestCase):

    def setUp(self):
        super(HelloWorld, self).setUp()
        self.app = self.start_app("Text Editor")

    def test_type_hello_world(self):
        self.keyboard.type("Hello World")


If you've used other testing frameworks that follow in the line of xUnit, you will notice the similarities. We implement an AutopilotTestCase object (class HelloWorld(AutopilotTestCase)), and define a new method for each test (ie, test_type_hello_world). You will also notice the setUp method. This is called before each test is run by the testrunner. In this case, we're launching the "Text Editor" application before we run each test (self.start_app("Text Editor")). Finally our test (test_type_hello_world) is simply sending keystrokes to type out "Hello World".

From this basic shell we can add more testcases to the helloworld testsuite easily by adding a new method. Let's add some simple ones now to show off some other capabilities of autopilot to control the mouse and keyboard. If you branched the bzr branch, there is a few more tests in the test_example.py file. These demonstrate some of the utility methods AutopilotTestCase makes available to us. Try running them now. The comments inside the file also explain briefly what each method does.

autopilot run helloworld.test_example.ExampleFunctions.test_keyboard
autopilot run helloworld.test_example.ExampleFunctions.test_mouse

Now there is more that autopilot can do, but armed with this basic knowledge we can put the final piece of the puzzle together. Let's create some assertions, or things that must be true in order for the test to pass. Here's a testcase showing some basic assertions.

autopilot run helloworld.test_example.ExampleFunctions.test_assert
  
Finally, there's some standards that are important to know when using autopilot. You'll notice a few things about each testsuite.
  • We have a folder named testsuite.
  • Inside the folder, we have a file named test_testsuite.py
  • Inside the file, we have TestSuite class, with test_testcase_name
  • Finally, in order for autopilot to see our testsuite we need to let python know there is a submodule in the directory. Ignoring the geekspeak, we need an __init__.py file (this can be blank if not otherwise needed)
Given the knowledge we've just acquired, we can tackle our first testcase conversion! For those of you who like to work ahead, you can already see the conversion inside the "firefox" folder. But the details, my dear Watson, will be revealed in due time. Until the next post, cheerio!