[Qgis-developer] Re: Renderer improvements masked for mac

[John C. Tull]

Hi Marco,

On Jan 3, 2010, at 1:15 AM, Marco Hugentobler wrote:

> Hi John
> 
> The incremental update was disabled because of update problems when resizing 
> the main window of QGIS on Mac. I tested it on a Mac at university and as far 
> as I remember, there was a difference with painting events coming from Qt that 
> caused the problems.
> 
> Could you test if this still is true? For testing, take out the ifdefs in 
> qgsoptions.cpp and also in qgsvectorlayer.cpp. Then take a large vector layer 
> that is updated several times. Start drawing and while it draws, resize the 
> QGIS window several times. Then do the same also with WMS layers.

Apparently I only tested the raster-related portion of the fix; I was not aware of the code in qgsvectorlayer.cpp. Removing that code does create some problems with qt-4.5.3. I'm going to download and try qt-4.6.0 and will report back on that if it works -- ok, tested that and it does not work any differently with 4.6.0.

> If there is no problem with it any more, I will be happy to remove the ifdefs, 
> as I like the incremental render updates very much too.

I suggest that we remove the mask in qgsoptions.cpp because it brings the raster speedups and there is no apparent cost to other functions. The raster caching makes a huge difference for me. I cannot comment on the quality of the vector improvements as I have not really tested it yet.

Cheers,
John

> Regards,
> Marco
> 
> Am Sonntag, 3. Januar 2010 02.49:42 schrieb John C. Tull:
>> Marco,
>> 
>> In trying to test the polygon dashed line issue, I removed the mask of the
>> following lines from qgsoption.cpp that I added a few days ago on my
>> source file. The lines are:
>> 
>> #ifdef Q_WS_MAC //MH: disable incremental update on Mac for now to avoid
>> problems with resizing groupBox_5->setEnabled( false );
>> #endif //Q_WS_MAC
>> 
>> I am not clear why the above is set for Mac's. I'm not sure what the
>> resizing issue is. I see no ill effects when I mask these lines, and the
>> performance gains are much too great to not have this feature available.
>> This is truly makes qgis a much more enjoyable program for production
>> work, in my opinion.
>> 
>> Can we take this out unless there is a specific issue that can be pointed
>> out and tested?
>> 
>> BTW, this had no effect on the polygon boundary. I'm in the dark on why the
>> problem was present, but now is gone. Strange...
>> 
>> Regards,
>> John
>> 
> 
> -- 
> Dr. Marco Hugentobler
> HUGIS - GIS programming and consulting
> Webereistr. 66
> CH-8134 Zürich
> marco at hugis.net
> http://homepage.hispeed.ch/hugis/
> Technical Advisor QGIS Project Steering Committee