[Qgis-developer] Status of the multi-threaded rendering project

Andreas Neumann a.neumann at carto.net
Wed Jan 22 12:26:56 PST 2014

Hi Martin,

Thanks for the update. I am trying the threaded version from time to
time. Not yet for editing, but for using it as a viewer. I should invest
more time into trying it with editing features. I primarily tested
SpatiaLite and Postgis, with occasionally some rasters/WMS.

I agree that it would probably be best to integrate it into master after
2.2 - given the amount of changes and not enough testing yet.


Am 22.01.2014 17:10, schrieb Martin Dobias:
> Hi everyone
> As the feature freeze is approaching, I guess this is the right time
> for an update about the progress of multi-threded rendering.
> Since the last status update in December I have continued with the
> implementation, adding support for more data providers, fixing bugs
> and re-enabling features I disabled during the implementation [1]. The
> code is in a state where it is quite ready for the merge to master and
> I am satisfied with the outcome so far. Still, I believe it will be
> better not to rush into the merge now for 2.2 and then run into
> problems close to the release (or after!). Don't get me wrong - it
> works well for me and for the other few brave souls who helped with
> testing - but given the amount of changes it is likely that some
> problems will appear sooner or later... git diff statistics say: 315
> files changed, 13808 insertions(+), 9321 deletions(-).
> So, I suggest that we merge the threading branch as soon as the
> feature freeze will be lifted again, so we will have plenty of time
> until 2.4 release (4 months) for fixing problems that may appear.
> To get an idea of what is still missing, here's is my brief to-do list:
> - merge from master and resolve any conflicts (mainly geometry simplification)
> - update data providers (so that they work again):
>   - mssql (Nathan has kindly volunteered to get it done)
>   - sqlanywhere (needs to be done by folks who contributed it)
> - (nice to have) - ability to cancel raster block requests. Otherwise,
> rendering of raster layers may make the GUI unresponsive until they
> finish. GDAL currently does not support that either, so this may need
> also contributing to GDAL, otherwise only native raster providers
> (wms, wcs, grass) would benefit from that
> I would appreciate your observations from the testing!
> Regards
> Martin
> [1] https://github.com/wonder-sk/QGIS/commits/threading-revival
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer

More information about the Qgis-developer mailing list