These are the symbol layer classes where there's no special logic
required relating to feature rendering and features are rendered
one-by-one, with no sub symbols.
Optimise the logic used when the new geometry backend for
selective masking is in effect:
Whenever its SAFE, instead of calculating an "entire map" clipping
path and then applying this for every feature being rendered,
we now defer the calculation of the clipping path until we
are rendering individual features. Then, we create a clipping path
which contains ONLY the mask paths which are within the area
being drawn over.
This avoids having the entire map clipping path being used for
EVERY feature being rendered, which results in huge PDF/SVG
exports when masks are in effect, and instead results in
clipping paths which are confined just to a sensible area
around each rendered feature.
In some complex test projects this reduces the PDF export
size by a factor of 0.01!! (and results in PDFs/SVGs which
open much quicker in viewers and editors, and don't grind
their operation to a halt).
The original workaround was only supposed to apply to QList/
QVector of QVariantMaps. By replacing them all in sip we
break mapping of signals defined in c++ which are emitted
by Python code.
Fixes exceptions after running processing algorithms in Qt6
builds.
And make sipify handle this nicely. This means that all our non-flag
style enums correctly map across to IntFlag python enums on Qt 6,
fixing issues with negative enum values for these and providing
a better match for the original c++ enum.
Populating the style library from the database can be a time
consuming process, especially if a user has many symbols
present in their library.
But for many standalone scripts, and for qgis_process, the
style database may not be required.
Let's defer initialization of it until it's actually required,
saving the startup cost in qgis_process and 3rd party scripts.
On my system with a style database containing ~700 items this
cuts down qgis_process startup times by around 25%
