On Mon, Oct 21, 2013 at 06:05:07PM +0200, Alexander Sack wrote: > On Mon, Oct 21, 2013 at 5:12 PM, Olivier Tilloy > <olivier.til...@canonical.com> wrote: > > On Mon, Oct 21, 2013 at 4:55 PM, Jamie Strandboge <ja...@canonical.com> > > wrote:
> >> On 10/21/2013 05:58 AM, Olivier Tilloy wrote: > >> > > >> > > >> > > >> > On Mon, Oct 21, 2013 at 12:46 PM, Omer Akram <omer.ak...@canonical.com > >> > <mailto:omer.ak...@canonical.com>> wrote: > >> > > >> > Hello! > >> > > >> > webbrowser-app-autopilot depends on ubuntu-ui-toolkit-autopilot > >> > which > >> > provides reusable emulators(helpers) to easily interact with > >> > different > >> > components of the ubuntu-ui-toolkit while writing autopilot tests. > >> > The > >> > dependencies come from there. Many other apps depend > >> > on ubuntu-ui-toolkit-autopilot as well. > >> > > >> > > >> > That’s correct. And I should add that most applications’ autopilot tests > >> > should > >> > use the standard emulators provided by the SDK, and thus their > >> > -autopilot > >> > package should depend on ubuntu-ui-toolkit-autopilot. > >> > > >> > > >> > Some packaging expert will need to look at > >> > ubuntu-ui-toolkit-autopilot and > >> > see if we can do some packaging changes there to reduce the deps. > >> > > >> > > >> > Note that ubuntu-ui-toolkit-autopilot depends on dpkg-dev, which pulls > >> > in a lot > >> > of build dependencies, such as make and perl. This is because the > >> > emulators rely > >> > on dpkg-architecture to locate qmlscene. > >> > Maybe there’s a simpler way to achieve this, and we could get rid of the > >> > dependency on dpkg-dev? > >> > > >> > >> If dpkg-architecture is the only thing pulling in dpkg-dev, just use this > >> instead: > >> dpkg --print-architecture > > > > > > What the emulators do to determine where the qmlscene binary is is to > > invoke: > > > > dpkg-architecture -qDEB_HOST_MULTIARCH > > > > (see > > http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/tests/autopilot/ubuntuuitoolkit/base.py#L28). > > > > This is different from what `dpkg --print-architecture` returns. > > > Unfortunately, this doesn't give us the tuple: > > asac@thinki:/tmp$ dpkg --print-architecture > i386 > asac@thinki:/tmp$ dpkg-architecture -qDEB_HOST_MULTIARCH > i386-linux-gnu > Steve: I think we talked about a similar case a few weeks back? Did we > find a better way for them to do it right back then? Do you remember > that case? I think the last case we discussed was one where we concluded that the files being accessed shouldn't be in the multiarch-qualified path at all. For this one, the comment in the test is: """Return the command to launch qmlscene for autopilot tests.""" # We need to specify qt5 because qtchooser doesn't have a default # configuration on devices and it seems the environment variable # QT_SELECT=qt5 doesn't work for autopilot tests. --Mirv - 2013-10-03 Maybe we could get to the bottom of why that's not working, instead of bypassing the interface? This seems to work fine from the commandline. I also don't see any good reason for the Qt binaries to be installed to /usr/lib/$arch/qt5/bin instead of to /usr/lib/qt5/bin, but I realize that may be quite a bit more difficult to address given that Qt itself may expect binaries and libraries to all be relative to the same install root. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slanga...@ubuntu.com vor...@debian.org
signature.asc
Description: Digital signature
-- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
