[Marble-devel] Review Request: Clean gpsd position provider shutdown

Bastian Holst bastianholst at gmx.de
Wed Apr 28 08:58:13 CEST 2010



> On 2010-04-28 06:57:26, Bastian Holst wrote:
> > I also had no possibility, but it looks good to me.

Ah, no possibility to test it.


- Bastian


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/3794/#review5276
-----------------------------------------------------------


On 2010-04-23 19:38:29, Dennis Nienhüser wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/3794/
> -----------------------------------------------------------
> 
> (Updated 2010-04-23 19:38:29)
> 
> 
> Review request for marble.
> 
> 
> Summary
> -------
> 
> When gpsd is running, but no data is available, the gpsd position provider is currently blocked indefinitely for libgsp >= 2.90 by the poll() call. In that case the destruction of the thread hangs in wait() after terminate(), which leads to marble freezing when turning gpsd off. It may also cause the crash reported in https://bugs.kde.org/show_bug.cgi?id=234311.
> 
> The patch changes the behavior to only call the blocking method when data is available (and it won't block in that case). Additionally it removes the possibly dangerous (according to qt docs) terminate() call.
> 
> 
> This addresses bug 234311.
>     https://bugs.kde.org/show_bug.cgi?id=234311
> 
> 
> Diffs
> -----
> 
>   /trunk/KDE/kdeedu/marble/src/plugins/positionprovider/gpsd/GpsdConnection.cpp 1117994 
>   /trunk/KDE/kdeedu/marble/src/plugins/positionprovider/gpsd/GpsdPositionProviderPlugin.cpp 1117994 
> 
> Diff: http://reviewboard.kde.org/r/3794/diff
> 
> 
> Testing
> -------
> 
> Using libgps-2.92-4 and some gps device supported by gpsd.
> 
> 
> Thanks,
> 
> Dennis
> 
>



More information about the Marble-devel mailing list