Re: [gdal-dev] Use of std::unique_ptr and auto in new code

2017-12-12 Thread Kurt Schwehr
+1 on the range based for +1 on std::unique_ptr, but ... I think GDAL should stick to these: - Do not do "using std::unique_ptr;" - Do not exposed std::unique_ptr to any public interfaces (for the time being) On Tue, Dec 12, 2017 at 4:06 AM, Even Rouault wrote: > >

[gdal-dev] Sub-optimal access to cloud optimized GeoTIFF

2017-12-12 Thread Rahkonen Jukka (MML)
Hi, This image http://193.166.24.122/gtk/mp_50_750.tif should be prepared as adviced in https://trac.osgeo.org/gdal/wiki/CloudOptimizedGeoTIFF and the validation script agrees on that. python c:\temp\validate_cloud_optimized_geotiff.py mp_50_750.tif Result:

Re: [gdal-dev] Use of std::unique_ptr and auto in new code

2017-12-12 Thread Even Rouault
> 1. std::unique_ptr > [...] > > Hopefully this will help to reduce the incidence of memory leaks. +1, that's one of the benefits of C++11 > > -- > > 2. auto > > I'd like to not have to write > > for( std::vector::iterator oIt = >apoMyList.begin(); > > and simply

[gdal-dev] Use of std::unique_ptr and auto in new code

2017-12-12 Thread Alan Thomas
Following up on Kurt's email: > It would be worthwhile to (even if only briefly) discuss the use each > new feature before starting to committing changes. I would like to begin to use the following C++11 features in new code: -- 1. std::unique_ptr Up until now, the DXF driver uses containers