[Marble-devel] Review Request: Fix runners on Qt 4.8

[Bernhard Beschow]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/103711/
-----------------------------------------------------------

(Updated Jan. 17, 2012, 12:21 a.m.)


Review request for Marble.


Changes
-------

testing++


Description
-------

Instantiating a QEventLoop is now subject to the runners. Only runners which use asynchronous API (networking) make use of QEventLoops, while the others perform their work in the respective methods (which are run in a thread).

In addition, the runners are now instantiated in the background thread rather than in the GUI thread, which avoids thread affinity issues. In particular, it seems to avoid QEventLoops to be assigned to the GUI thread, causing them to block forever in the background threads.


This addresses bug 288612.
    http://bugs.kde.org/show_bug.cgi?id=288612


Diffs
-----

  src/lib/MarbleRunnerManager.h 18c0864 
  src/lib/MarbleRunnerManager.cpp 9b04537 
  src/lib/RunnerTask.h 6b4f7aa 
  src/lib/RunnerTask.cpp fb4d1a5 
  src/plugins/runner/hostip/HostipRunner.h b4bdd37 
  src/plugins/runner/hostip/HostipRunner.cpp f079f5d 
  src/plugins/runner/nominatim/OsmNominatimRunner.cpp f55975c 
  src/plugins/runner/openrouteservice/OpenRouteServiceRunner.cpp 015381d 
  src/plugins/runner/yours/YoursRunner.cpp a86a34e 
  tests/MarbleRunnerManagerTest.cpp 3f2fb10 

Diff: http://git.reviewboard.kde.org/r/103711/diff/diff


Testing (updated)
-------

online + offline routes, addresses, and online + offline placemarks were found and displayed successfully on the following platforms:
* x86 desktop with Qt 4.8.0
* Maemo 5 (N900) with Qt 4.7.0
* desktop Windows (wine) with Qt 4.7.4


Thanks,

Bernhard Beschow

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/marble-devel/attachments/20120117/04681d1a/attachment.html>