lp:~ev/ubuntu/maverick/jockey/auto_install

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp:ubuntu/maverick/jockey

Martin Pitt: Approve on 2010-07-21

Registry Administrators: Pending requested 2010-07-16

Diff: 173 lines (+74/-0)

8 files modified

data/handlers/broadcom_wl.py (+1/-0)
debian/changelog (+6/-0)
jockey/backend.py (+1/-0)
jockey/handlers.py (+6/-0)
jockey/ui.py (+13/-0)
tests/handlers.py (+12/-0)
tests/sandbox.py (+11/-0)
tests/ui.py (+24/-0)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

87. By Evan on 2010-07-20

Add test cases.

86. By Evan on 2010-07-16

Add an --auto-install option.

85. By Jonathan Thomas on 2010-05-16

Remove the X-KDE-SubstitueUID key from jockey-kde.desktop.in. It is no
longer needed since Jockey now has PolicyKit support, and poses a less
than desirable situation where the whole frontend is run as root.
Fixes LP: #581379

84. By Alberto Milone on 2010-04-19

* data/handlers/fglrx.py, nvidia.py:
  - Do not check the non null value of current_alternative and
    target_alternative. This caused jockey to skip the condition
    that we really wanted to test (i.e. do the two alternatives
    match?) and show the driver as enabled when the target
    alternative didn't exist (as the package hadn't been installed
    yet) (LP: #562226).

83. By Alberto Milone on 2010-04-12

[ Alberto Milone ]
* Do not check that the module is loaded in enabled() in the nvidia
  handler, as this is what we should do in used(). Just make sure
  that the module and its alias are not blacklisted. This prevents
  Jockey from claiming that the installation of nvidia driver failed
  when when it didn't (LP: #552653).

[ Jonathan Thomas ]
* Don't crash when closed with the "x" button (LP: #552723, LP: #398912).
  Thanks ThorbjørnTux!

[ Martin Pitt ]
* i18n "Install Drivers" string. (LP: #542552)

82. By Alberto Milone on 2010-03-30

* data/handlers/fglrx.py:
  - Re-enable the fglrx handler and make it use the alternatives system
    (LP: #496225).
* data/handlers/nvidia.py:
  - See if the module alias is loaded and then check what module it
    corresponds to in order to see if it's really enabled (LP: #547066).
    This involved overriding the enabled() and the used() methods.
* jockey/xorg_driver.py:
  - Simplify the logic in the enabled() method by only relying on
    X-Kit's isDriverEnabled() method. This will avoid getting false
    positives in the case in which the 1st Device section (i.e. the
    only one that we used to check) uses the right driver but the
    same section is irrelevant as it's not mentioned in the
    ServerLayout section.

81. By Martin Pitt on 2010-03-26

* Merge fixed KDE .ui file from trunk, thanks Jonathan Thomas! This fixes
  the crash on jockey-kde startup.
* debian/control: Revert the python-distutils-extra build dependency bump,
  not necessary any more with the fix above.

80. By Martin Pitt on 2010-03-26

* Merge bug fix from trunk:
  - ui.py: Fix _check_repositories() to not invoke the UI and error dialogs
    in --check mode. (LP: #540750)
* debian/control: Bump python-distutils-extra build dependency, to ensure we
  build correct -kde packages. (LP: #543707)

79. By Mario Limonciello on 2010-03-25

[ Martin Pitt ]
* jockey/oslib.py: Use "dkms" instead of "bash" for testing whether we have
  repositories. This works slightly better with third-party repositories.

[ Mario Limonciello ]
* data/handlers/fglrx.py: The package name is now just "fglrx" rather than
  xorg-driver-fglrx.

78. By Martin Pitt on 2010-03-20

* Merge fixes from trunk:
  - ui.py, check(): If the user is running Jockey with package installation
    or another long-running task in parallel, the automatic --check
    invocation will time out on connecting to the backend. This is not
    fatal, just print an error message instead of crashing. (LP: #403955)
  - jockey-gtk: Fix progress dialogs to not be maximizable. (LP: #537911)
  - Do not parent the KMessageBox, since in some cases the errors can come
    before the MainWindow is initialized. Thanks Jonathan Thomas!