lp:~zyga/plainbox-provider-phablet/flags

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp:plainbox-provider-phablet at revision 10

Daniel Manrique (community): Approve on 2014-07-14

Diff: 79 lines (+9/-2)

2 files modified

2013.com.canonical.certification:phablet/units/graphics.txt (+4/-2)
2013.com.canonical.certification:phablet/units/radio.txt (+5/-0)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

10. By Zygmunt Krynicki on 2014-07-14

jobs: flag wireless-002 as a multi-node test

The multi-node term originates from my experience with LAVA, where LAVA
went through similar evolution that we are going through now. Initially
starting with just running on-device (lava-test), then getting a way
to run remotely from a single "host", executing code on a single "target"
(lava-dispatcher) to finally, being able to coordinate and execute tests
on a set of heterogeneous machines (mapping tests to machines in a free
way).

I believe that this is an instance of that type of testing and it would
require more work, inside plainbox, to support. Here the test session
would be comprised of two machines, the local machine and one phablet
machine. Only in that arrangement the test makes sense. What is worse for
us is that if it is included in a larger test collection (whitelist)
plainbox will need to figure out which tests need to execute on which
devices and in what order. That is perfectly achievable but it is a sprint
in itself, comparable in size to our recent "remoting" sprint in Budapest.

That also has implications on any UI capable of executing such tests
in complex environments.

So all in all, this test is okay but will probably need to change
to be a normal remote test, without relying on running anything on the
host. I'd like to finish remote support across the stack before exploring
multi-node support that would make jobs like this possible.

Signed-off-by: Zygmunt Krynicki <email address hidden>

9. By Zygmunt Krynicki on 2014-07-14

jobs: flag jobs using adb_test with 'legacy-remote' flag

This patch is meant to let us track some jobs that currently rely on the
adb_test script, which must run on the host, so that as plainbox gets
native remote execution support in trunk, we can convert those to run
remotely with 'plainbox run -@ phablet'.

Currently nothing looks at this flag, perhaps as we go forward we'll get
a website (hint hint) that displays providers :-)

Signed-off-by: Zygmunt Krynicki <email address hidden>

8. By Nara Huang on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=narahuang]"

7. By Jonathan Cave on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=jocave]"

6. By Nara Huang on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=narahuang]"

5. By Jonathan Cave on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=jocave]"

4. By Jonathan Cave on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=jocave]"

3. By Nara Huang on 2014-06-26

"automatic merge by tarmac [r=roadmr][bug=][author=narahuang]"

2. By Daniel Manrique on 2014-06-26

"Added per-commit testing harness.

This adds the tools needed to do testing with vagrant or lxc, as well as a tarmac-verify script. [r=zkrynicki][bug=][author=roadmr]"

1. By Daniel Manrique on 2014-06-25

First commit - lots of boilerplate code here. Watch out.