Previously pictures could only be synced to grid north, which
can be totally wrong for many CRSes (especially in polar areas)
Users now are given a choice of grid or true north, and can also
enter an optional offset to apply if eg magnetic north is instead
desired.
When synced to true north the bearing is calculated using the
centre point of the linked map item.
Fix#192, #4711
This fix was sponsored by the Norwegian Polar Institute's
Quantarctica project (http://quantarctica.npolar.no) and
coordinated by Faunalia.
Adds a new checkbox at the bottom of each symbol layer's
properties which allows you to control whether the layer is
enabled or not.
Disabled layers are not drawn, but are saved and can be
enabled at a later stage.
This makes it easier to tweak symbol appearance without
having to totally delete a symbol layer.
When saving a vector layer into an existing file, depending on the capabilities
of the output driver, the user can now decide whether:
- to overwrite the whole file
- to overwrite only the target layer (layer name is now configurable)
- to append features to the existing target layer
- to append features, add new fields if there are any.
All above is available for drivers like GPKG, SpatiaLite, FileGDB, ...
For drivers like Shapefile, MapInfo .tab, feature append is also available.
The order of the elements is irrelevant and duplicate elements are unwanted. It
is therefore a perfect candidate for a set instead of a list. This prevents
filtering for duplicates manually be replacing some filer codes with (more
performant) builtin methods of QSet.
Groups nearby points into a single rendered marker symbol.
QgsPointDisplacementRenderer has been split into a new
pure virtual QgsPointDistanceRenderer base class which handles the
detection of clusters and grouping of points. The new cluster
renderer reuses this base class to avoid code duplication.
Additionally, some improvements have been made to the
displacement renderer, specifically:
- points are now assigned to the group which is "nearest" them,
rather then just assigning them first group within the search
distance. In some cases this was assigning features to a more
distant cluster, resulting in less predictable cluster patterns
- individual points are now correctly shown in their own
selection state
Lots of code cleanup + documentation too.
Sponsored by:
- Andreas Neumann
- Qtibia Engineering (Tudor Barascu)
- Karl-Magnus Jönsson
- Geonesia (Nicolas Ponzo)
- Plus numerous additional anonymous backers whose generous
contributions are also highly valued!
Fixed the add relation functionnality: the table is sorted. When the code
was setting the sorted column, the row was sorted and the other columns it was
setting were set on the wrong row.
