When a topological check returns a gap, it's sometimes better to zoom to the gap,
sometimes to show the gap in the context of the surrounding polygons. There is no
one-size-suits-all solution.
Therefore it's now possible to zoom to one by enabling the "zoom to feature(s)" button
also for topological errors.
The geometry validation only works on the current edit session (added / edited geometries). To detect topology
errors it is required to also get more features within the context, therefore, the bounding box of the edited
geometries is taken to populate the list of features to check.
This commit filters the found problems so only the ones which actually affect one of the edited geometries
will be reported.
For the c++ api dox this expands to "\c nullptr" (the
\c directive indicates a code literal value), and for sipify/Python
it expands to ``None`` (`` is sphinx annotation for literal values)
Makes for nicer dox for both c++ and Python!
If they are deleted, any other python wrappers which have been parented to these singletons are deleted too.
One of the most common issues with this is, that subclasses of registry items are deleted and lost.
Similarly to raster calculator, mesh calculator can take dataset groups from current mesh layer and
combine them with various aritmentic/logical operators to new dataset group.
Contains utilities for retrieving the geotag from images and
for setting an image's geotag.
Working with geotags (before this class!) is super-annoying
and fiddly and relies on either parsing command line tools
or depending on non-standard Python libraries which are
not available everywhere, and often very difficult for users
on certain platforms to get installed and working correctly.
With this class we have stable methods for geotag getting/setting
which are universally available and can be used safely
by plugins and scripts.
Fix a deadlock in geometry validation. This happens when the mainthread "waits for finished" of a checker thread.
What happened in this case was, that the main thread canceled the feedback and waited for any jobs to finish.
If a job was waiting for a feature source (or something else to be executed on the main thread) this resulted in a mighty deadlock.
What we do here is that we regularly check if we ought to cancel while waiting for our slot on the main thread and can bail out if we should cancel before the main thread gets around to take care of our function.
The difference is, we still execute the code on the background thread and make sure that the main thread is not doing anything during this time to avoid working on the same data structures in parallel.
