A test failing on builders with mir 0.1.1: PublishedSocketConnector.drm_auth_magic_is_processed_by_the_server

Yes, the armhf tests aren't meant to run yet. We have some bugs to fix in them before they will work.

I can't reproduce the failure either :( Obviously the tests work for all of us or else it would not have been released like that.

You could just disable the failing test for now, by changing:
TEST_F(PublishedSocketConnector, drm_auth_magic_is_processed_by_the_server)
to
TEST_F(PublishedSocketConnector, DISABLED_drm_auth_magic_is_processed_by_the_server)
[tests/unit-tests/frontend/test_published_socket_connector.cpp]

It's likely just a problem in the test with headless environments, or something like that.

Actually it's likely just the timeout is too short (0.1s):
        client = std::make_shared<mt::TestProtobufClient>(test_socket, 100);

Annoying while it built in the qt5-daily PPA I used for testing, it's still failing in the daily-build PPA with the branch merged:

https://launchpad.net/~ubuntu-unity/+archive/daily-build/+sourcepub/3667324/+listing-archive-extra

The successful PPA build was https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-daily/+sourcepub/3666854/+listing-archive-extra

(the successful build has the change, just not the changelog entry)

The test is now segfaulting on i386 and amd64...

See https://launchpadlibrarian.net/156840778/buildlog_ubuntu-trusty-amd64.mir_0.1.1%2B14.04.20131118.1-0ubuntu1_FAILEDTOBUILD.txt.gz
and
https://launchpadlibrarian.net/156840798/buildlog_ubuntu-trusty-i386.mir_0.1.1%2B14.04.20131118.1-0ubuntu1_FAILEDTOBUILD.txt.gz

Fix committed into lp:~mir-team/mir/development-branch at revision 1234, scheduled for release in mir, milestone Unknown

Fix committed into lp:mir at revision 1165, scheduled for release in mir, milestone 0.1.2

This bug was fixed in the package mir - 0.1.1+14.04.20131120-0ubuntu1

---------------
mir (0.1.1+14.04.20131120-0ubuntu1) trusty; urgency=low

  [ Daniel van Vugt ]
  * New upstream release 0.1.1
    - Add unit tests for V/H scroll events.
    - surfaces: avoid publishing some internal headers, tidy up default
      configuration, integrate surfaces report.
    - client: Add mir_connection_drm_set_gbm_device()
    - graphics: avoid publishing some internal headers.
    - Fixed: unity-system-compositor FTBFS on trusty against new Mir
      (libmirserver9) (LP: #1244192)
    - compositor: avoid publishing some internal headers.
    - shell: Add set_lifecycle_state() to the Session interface.
    - frontend: avoid publishing some internal headers
    - logging: avoid publishing some internal headers.
    - Allow specifying the nested server name by passing --name= or setting
      MIR_SERVER_NAME=.
    - graphics,gbm: Inform the EGL platform about the used gbm device when
      using the native GBM platform
    - examples: Restore GL state after initializing buffers, fixing crashes
      observed in render_surfaces (LP: #1234563)
    - Continue refactoring the mir android display classes.
    - shell: Hoist focus control functions needed by unity-mir into
      FocusController interface
    - client: Remove the timeout for detecting server crashes
    - Avoid a race condition that could lead to spurious failures of server
      shutdown tests (LP: #1245336)
    - test_client_input.cpp: Bump reception time-out in client test fixture.
      (LP: #1227683)
    - Ensure StubBufferAllocator returns buffers with the properties requested,
      and not the same old hardcoded constants.
    - Update docs and scripting for trusty.
    - compositor: Make DefaultDisplayBufferCompositorFactory private to the
      compositor component.
    - Ignore warnings clang treats as errors, about unused functions being
      generated from macros in <lttng/tracepoint.h> (LP: #1246590)
    - Add resize() support to BufferBundle. This is the first step and lowest
      level of surface resize support.
    - Clean up constants relating to SwitchingBundle.
    - Fix the armhf chroot setup script to point to the right library, so
      cross compiling can work again (LP: #1246975)
    - shell: avoid publishing some internal headers.
    - input: avoid publishing some internal headers.
  * Bump timeouts used in socket testing. It seems 100ms isn't always
    enough, which leads to spurious test failures (LP: #1252144) (LP:
    #1252144)
  * Fix uninitialized variable causing random drm_auth_magic test
    failures. (LP: #1252144). (LP: #1252144)

  [ Ubuntu daily release ]
  * Automatic snapshot from revision 1165
 -- Ubuntu daily release <email address hidden> Wed, 20 Nov 2013 07:36:15 +0000