2ac5933 fixed the regression in 2.12, but there were more underlying
issues from <2.8 causing sub symbols with data defined properties
to be ignored.
Add some tests.
(refs #13707)
This introduces two new options to filter legend elements:
- filter by expression: a boolean expression can be set. Only symbols of
features that make the expression evaluated to true will be kept in the legend
- filter by polygon: only symbols of features that are inside the given
polygon will be part of the legend.
The polygon filtering is used in particular for a new option in the
composer legend that allows to filter out anything that is not included
in the current atlas polygon.
Not perfect, but good for a quick warning if a new class or member
has been added to the public API without Python bindings. The test
only considers the name of members, since it seems to be impossible
to test for the signature of a Python member. (So adding a new
overloaded method without bindings will still unfortunately pass).
You can avoid the test where bindings are not applicable:
- for a whole class by placing "@note not available in Python bindings"
in the class' Doxygen comments
- or by placing the @note inside a member's Doxygen comments for a
specific member
Additionally, classes which aren't included in the API docs
will not be tested.
- when creating geometry from WKT, upgrade dimensionality of geometry
if coordinates are 3/4 dimensional
- match dimensionality of collections to child dimensionality
- fix area of curves was non-zero if curve is closed
- don't consider m values when testing for curve closedness
- add unit tests for closedness
- add unit tests for CircularStrings, CompoundCurves, CurvePolygon,
tests with geometries with Z/M values
Split length from perimeter calculation in geometry API, as
returning perimeter for length for polygons is misleading and
results in incorrect length/perimeter calculations for mixed
geometry collections.
Enable length & perimeter unit tests against reference geometries.
Now the length & perimeter values match those calculated by
PostGIS.
- Add some tests for conversion to/from WKT, using a bulk lot of testsdata
from PostGIS
- Add some tests for area/length calculation, using some test data and
results from PostGIS/geos unit tests
- Add tests for spatial relations, using test data from PostGIS. Note
that this required adding support for calculating the DE-9IM relation. I'll
expose this to users via the expression engine in 2.14.
Along the way this also fixes a number of bugs relating to WKT geometry
import, such as
- add support for alternate MultiPoint(1 1,2 2,...) format
- fix GeometryCollection to support collections with multi* children
and GeometryCollection children (allowed by spec)
I'm not sure what sip voodoo is preventing the bindings from
calling methods in the base class, so for QGIS <3.0 I've aliased
all the new signatures and will rename them back for 3.0
* fieldNameIndex: resort to case-insensitive lookup only if
case-sensitive lookup has no match
* file writer: fix handling of fields that only differ by case
reset the fill/border color and border width when changing svg
marker/svg fill SVG files
This change makes the behaviour consistent between the svg marker
symbol and the other marker symbols. Additionally, svg files
which have customisable colors and NO default values set will
be shaded in gray fill/black outline in the svg selector widget,
to follow the same behaviour as the other marker symbol selectors.
Note that this change has NO EFFECT unless the svg files are
modified to remove the default param value, so there will be no
change for users' custom symbols. A follow up commit will need
to remove the default param values from the preinstalled SVG
files though. If you want to test in the meantime, I've modified
just the first two symbols in the accomodation group to make
this change for testing.
(refs #10908)
Rationale:
- there was a lot of large objects passed by value, so potentially
there's a speed bump from this
- even for implicitly shared classes like QString/QList there's still
a (small) cost for copying the objects when there's no reason to
- it's the right thing to do!
- call reserve on container classes where applicable
- make sure Q_FOREACH uses references
- remove dynamic casts to base classes
- fix some implicit bool conversions
