Artikel getaggt mit announcement

autopkgtest 3.5: Reboot support, Perl/Ruby implicit tests

Last week’s autopkgtest 3.5 release (in Debian sid and Ubuntu Utopic) brings several new features which I’d like to announce.

Tests that reboot

For testing low-level packages like init or the kernel it is sometimes desirable to reboot the testbed in the middle of a test. For example, I added a new boot_and_services systemd autopkgtest which configures grub to boot with systemd as pid 1, reboots, and then checks that the most important services like lightdm, D-BUS, NetworkManager, and cron come up as expected. (This test will be expanded a lot in the future to cover other areas like the journal, logind, etc.)

In a testbed which supports rebooting (currently only QEMU) your test will now find an “autopkgtest-reboot” command which the test calls with an arbitrary “marker” string. autopkgtest will then reboot the testbed, save/restore any files it needs to (like the tests file tree or previously created artifacts), and then re-run the test with ADT_REBOOT_MARK=mymarker.

The new “Reboot during a test” section in README.package-tests explains this in detail with an example.

Implicit test metadata for similar packages

The Debian pkg-perl team recently discussed how to add package tests to the ~ 3.000 Perl packages. For most of these the test metadata looks pretty much the same, so they created a new pkg-perl-autopkgtest package which centralizes the logic. autopkgtest 3.5 now supports an implicit debian/tests/control control file to avoid having to modify several thousand packages with exactly the same file.

An initial run already looked quite promising, 65% of the packages pass their tests. There will be a few iterations to identify common failures and fix those in pkg-perl-autopkgtest and autopkgtestitself now.

There is still some discussion about how implicit test control files go together with the DEP-8 specification, as other runners like sadt do not support them yet. Most probably we’ll declare those packages XS-Testsuite: autopkgtest-pkg-perl instead of the usual autopkgtest.

In the same vein, Debian’s Ruby maintainer (Antonio Terceiro) added implicit test control support for Ruby packages. We haven’t done a mass test run with those yet, but their structure will probably look very similar.

Tags: , , , , , ,

autopkgtest 3.2: CLI cleanup, shell command tests, click improvements

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
--output-dir=/tmp/out
-s
---
schroot
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/README.click-tests.html 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!

Tags: , , , ,

deb, click, schroot, LXC, QEMU, phone, cloud: One autopkgtest to Rule Them All!

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 http://people.canonical.com/~pitti/tmp/com.ubuntu.calculator_1.3.283_all.click
$ adt-run ubuntu-calculator-app/ com.ubuntu.calculator_1.3.283_all.click --- \
      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/test_simple_page.py", line 93, in test_divide_with_infinity_length_result_number
    self._assert_result("0.33333333")
  File "/tmp/adt-run.KfY5bG/tree/tests/autopilot/ubuntu_calculator_app/tests/test_simple_page.py", line 63, in _assert_result
    self.main_view.get_result, Eventually(Equals(expected_result)))
  File "/usr/lib/python3/dist-packages/testtools/testcase.py", 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/ com.ubuntu.calculator_1.3.283_all.click --- 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/ com.ubuntu.calculator_1.3.283_all.click \
          --- 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!

Tags: , , , , , ,

What’s the autopilot widget that I want?

Today’s autopilot release provides a new feature for test case writers. Unless the widget you want to test has a direct object name (GtkBuilder ID/Qt objectName), it is often not that easy to find a widget in a deeply nested hierarchy in autopilot vis.

With the new version, if you have some parent widget (like the containing dialog) w in your test, you can now call w.print_tree() to dump the paths and properties of that widget and all its children to stdout. That’s easy enough to grep, so provides a “poor man’s full tree search”. You can also specify a different output sink, like a file object or a file name: w.print_tree('/tmp/dump.txt').

This is a first step towards making it easier to find widgets and properties you are interested in. Arguably this is mostly just a crutch, but I found it to be rather effective. Before this feature I often wrote little snippets like in LP#1241312, now this becomes much easier. A better solution for this would certainly be a “full tree search” in vis itself, but that’s not that easy to implement. It is on the roadmap for this cycle, though.

I am also currently working on a real-time property change monitor for autopilot-gtk, which may also help in some cases. Unfortunately we cannot build such a thing for autopilot-qt, as due to the nature of Qt object properties, changes of them cannot be monitored.

Tags: , , ,

umockdev 0.4: Mocking phone calls

umockdev 0.3 introduced the notion of an “umockdev script”, i. e. recording the read()s and write()s that happen on a device node such as ttyUSB0. With that one can successfully run ModemManager in an umockdev testbed to pretend that one has e. g. an USB 3G stick.

However, this didn’t yet apply to the Ubuntu phone stack, where ofonod talks to Android’s “rild” (Radio Interface Layer Daemon) through the Unix socket /dev/socket/rild. Thus over the last days I worked on extending umockdev’s script recording and replaying to Unix sockets as well (which behave quite different and quite a bit more complex than ordinary files and character devices). This is released in 0.4, however you should actually get 0.4.1 if you want to package it.

So you now can make a script from ofonod how it makes a phone call (or other telephony action) through rild, and later replay that in an umockdev testbed without having to have a SIM card, or even a phone. This should help with reproducing and testing bugs like ofonod goes crazy when roaming: It’s enough to record the communication for a person who is in a situation to reproduce the bug, then a developer can study what’s going wrong independent of harware and mobile networks.

How does it work? If you have used umockdev before, the pattern should be clear now: Start ofonod under umockdev-record and tell it to record the communication on /dev/socket/rild:

  sudo pkill ofonod; sudo umockdev-record -s /dev/socket/rild=phonecall.script -- ofonod -n -d

Now launch the phone app and make a call, send a SMS, or anything else you want to replay later. Press Control-C when you are done. After that you can run ofonod in a testbed with the mocked rild:

  sudo pkill ofonod; sudo umockdev-run -u /dev/socket/rild=phonecall.script -- ofonod -n -d

Note the new --unix-stream/-u option which will create /tmp/umockdev.XXXXXX/dev/socket/rild, attach some server threads to accept client connections, and replay the script on each connection.

But wait, that fails with some

   ERROR **: ScriptRunner op_write[/dev/socket/rild]: data mismatch; got block '...', expected block '...'

error! Apparently ofono’s messages are not 100% predictable/reproducible, I guess there are some time stamps or bits of uninitialized memory involved. Normally umockdev requires that the program under test sticks to the previously recorded write() parts of the script, to ensure that the echoed read()s stay in sync and everything works as expected. But for cases like these were some fuzz is expected, umockdev 0.4 introduces setting a “fuzz percentage” in scripts. To allow 5% byte value mismatches, i. e. in a block of n bytes there can be n*0.05 bytes which are different than the script, you’d put a line

  f 5 -

before the ‘w’ block that will get jitter, or just put it at the top of the file to allow it for all messages. Please see the script format documentation for details.

After doing that, ofonod works, and you can do the exact same operations that you recorded, with e. g. the phone app. Doing other operations will fail, of course.

As always, umockdev-run -u is of course just a CLI convenience wrapper around the umockdev API. If you want to do the replay in a C test suite, you can call

   umockdev_testbed_load_socket_script(testbed, "/dev/socket/rild",
                                       SOCK_STREAM, "path/to/phonecall.script", &error);

or the equivalent in Python or Vala, as usual.

If you are an Ubuntu phone developer and want to use this, please don’t hesitate to talk to me. This is all in saucy now, so on the Ubuntu phone it’s a mere “sudo apt-get install umockdev” away.

Tags: , , , , , , ,

umockdev 0.3: record and replay of tty devices

I’m happy to announce a new release 0.3 of umockdev.

The big new feature is the ability to fake character devices and provide recording and replaying of communications on them. This work is driven by our need to create automatic tests for the Ubuntu phone stack, i. e. pretending that we have a 3G or phone driver and ensuring that the higher level stacks behaves as expected without actually having to have a particular modem. I don’t currently have a phone capable of running Ubuntu, so I tested this against the standard ModemManager daemon which we use in the desktop. But the principle is the same, it’s “just” capturing and replaying read() and write() calls from/to a device node.

In principle it ought to work in just the same way for other device nodes than tty, e. g. input devices or DRI control; but that will require some slight tweaks in how the fake device nodes are set up; please let me know if you are intested in a particular use case (preferably as a bug report).

With just using the command line tools, this is how you would capture ModemManager’s talking to an USB 3G stick which creates /dev/ttyUSB{0,1,2}. The communication gets recorded into a text file, which umockdev calls “script” (yay my lack of imagination for names!):

# Dump the sysfs device and udev properties
$ umockdev-record /dev/ttyUSB* > huawei.umockdev

# Record the communication
$ umockdev-record -s /dev/ttyUSB0=0.script -s /dev/ttyUSB1=1.script \
     -s /dev/ttyUSB2=2.script -- modem-manager --debug

The –debug option for ModemManager is not necessary, but it’s nice to see what’s going on. Note that you should shut down the running system instance for that, or run this on a private D-BUS.

Now you can disconnect the stick (not necessary, just to clearly prove that the following does not actually talk to the stick), and replay in a test bed:

$ umockdev-run -d huawei.umockdev -s /dev/ttyUSB0=0.script -s /dev/ttyUSB1=1.script \
    -s /dev/ttyUSB2=2.script -- modem-manager --debug

Please note that the CLI options of umockdev-record and umockdev-run changed to be more consistent and fit the new features.

If you use the API, you can do the same with the new umockdev_testbed_load_script() method, which will spawn a thread that replays the script on the faked device node (which is just a PTY underneath).

If you want full control, you can also do all the communication from your test cases manually: umockdev_testbed_get_fd("/dev/mydevice") will give you a (bidirectional) file descriptor of the “master” end, so that whenever your program under test connects to /dev/mydevice you can directly talk to it and pretend that you are an actual device driver. You can look at the t_tty_data() test case for how this looks like (that’s the test for the Vala binding, but it works in just the same way in C or the GI bindings).

I’m sure that there are lots of open ends here still, but as usual this work is use case driven; so if you want to do something with this, please let me know and we can talk about fine-tuning this.

In other news, with this release you can also cleanly remove mocked devices (umockdev_testbed_remove_device()), a feature requested by the Mir developers. Finally there are a couple of bug fixes; see the release notes for details.

I’ll upload this to Saucy today. If you need it for earlier Ubuntu releases, you can have a look into my daily builds PPA.

Let’s test!

Tags: , , , , , ,

PyGObject 3.8.3 released

While GNOME as a whole does not have a planned 3.8.3 release, I got some requests to do a new stable release of PyGObject with some important bug fixes, so here it is: version 3.8.3. Thanks to all contributors!

  • Add marshalling of GI_TYPE_TAG_VOID held in a GValue to int. While not particularly useful this allows some callbacks in WebKit to function without causing a segfault. (Simon Feltman) (#694233)
  • pygtkcompat: Fix for missing methods on Windows (Martin Pitt) (#702787)
  • gi/pygi-info.c: Avoid C99-style variable declaration (Chun-wei Fan) (#702786)
  • Clear return value of closures to zero when an exception occures (Simon Feltman) (#702552)
  • Re-add support for passing GValue’s by reference (Simon Feltman) (#701058)
  • Don’t use doctest syntax in docstrings for examples, to fix test failures with pyflakes 0.7.x (Martin Pitt) (#701009)
  • examples/option.py: Port to GI and Python 3 (Martin Pitt)

Tags: , , , , ,

umockdev 0.2.6: Hello ARM

I released umockdev 0.2.6. Most importantly, this now fully works on ARM platforms, as we want to use it to write tests for/on the Ubuntu phone. I tested it on my Nexus 7, and the tests also succeed on the ARM Ubuntu builder (which are Panda boards). Fixing this revealed some interesting issues in recorded ioctl traces (as they are platform specific in some cases due to different word length) as well as kernel bugs in the Tegra drivers.

This version also fixes compatibility with older automake versions again, so that the daily builds for raring should work again.

I also have a new gvfs test case ready to commit which uses umockdev (if available) to test functionality of the gphoto backend. But that needs the new UMockdevTestbed.clear() API in 0.2.6, so I was holding that back. I will land it soon in upstream git now.

Tags: , , , , , ,

Ubuntu Saucy translations are now open

You can now start translating Ubuntu Saucy on Launchpad.

Tags: , , , ,

umockdev 0.2.2 released

I did a 0.2.2 maintenance release for umockdev to fix building with Vala 0.16.1, gcc 4.8 (the changed sizeof behaviour caused segfaults), and current udev releases (umockdev-record stumbled over the new “link priority” fields of udevadm). There are also a couple of bug fixes, but no new features.

Tags: , , , , ,