Instead of progress changed signals on the provider/layer level,
instead use QgsFeedback arguments for methods which previously
emitted progress changed signals
Previously the renderer required that pixel values followed
sequential numerical order. This refactor allows values to
be skipped, so that only certain color index will be rendered.
Instead of creating them for all items when composer window
is opened, instead just create and destroy them when
required. None are too heavy to have a noticable delay
when selecting items, but in contrast keeping them around
forever is wasteful on memory.
Also clean up a lot of duplicate composer/composition
code and remove unnecessary signals from api (use a single
itemAdded signal instead of multiple signals for every
item type)
Since composer windows are now only created on demand, and
destroyed when the window is closed, the old interface methods
no longer apply. The new interface methods openComposers(),
openComposer(), closeComposer(), composerOpened(),
composerWillBeClosed() and composerClosed() are similar, but
only apply to composer windows when they are exist (i.e.
are currently open).
To access all compositions from a project, the project's
layoutManager() should be used instead.
Additionally, the new interface methods work with
QgsComposerInterface objects instead of QgsComposerView
objects. This should allow future hooks for plugins to interact
with open composer windows in a more useful way.
This enum value has not been in use since filter rects were
separated from other filter types. Leaving it in API is confusing
and leads to incorrect use.
In a post-OTF-disabled world these are no longer required. The
map units are dictated by the map units for the destination
CRS.
It's dangerous to separate these two properties as setting
the incorrect map units for a given CRS will result in
incorrect scale calculations.
Because
- OTF reprojection is mature and stable
- Should be no cost when not required - transforms are shortcut
when not required
- Reduces code complexity
- Canvas OTF support was being incorrectly used as a flag for
whether measurements should be made in ellipsoidal mode. Instead
the project's ellipsoid setting should be checked for this.
- Should simplify behavior for new users
This change is being driven by:
- need for consistent terminology in the api. Currently there's
a confusing mix of border/outline used
- need for consistent terminology in the gui - we have a mix
of "border"/"outline"/"pen"
- bringing QGIS into line with the standard terminology used
in other graphic editing programs
A new class QgsPathResolver is introduced for conversion between absolute
and relative paths when reading/writing XML.
Cleaned up code in layer definition file support (.qlr) to better handle
relative/absolute path conversion.
Motivations:
- consistency - we generally use expanded names, and this also
matches Qt API which uses Database instead of Db
- avoids unpredictable capitalization throughout API (mix of "Db"
and "DB")
In some cases the render result for a layer may depend on more than
one layer.
This changes QgsMapRendererCache to use "cache keys" instead of layer ids
to identify images within the cache. When setting an image inside the
cache, a new list parameter is used to specify all layers on which
that image depends.
QgsMapRendererJob and subclasses are not designed to be subclassed
outside of core QGIS code. Marking them private API allows us to
change them after API is frozen again.
The basic idea is to split reading of XML definition and resolution of layer IDs into layers.
Also includes some cleanups in QgsProject and improvements in QgsVectorLayerJoinInfo.
Another bit in the project refactoring work to get rid of dependencies
on QgsProject singleton.
Reading of layer trees from XML is now split into two phases:
1. read XML and keep layer IDs
2. resolve layer IDs to QgsMapLayer instances (using QgsProject)
There are convenience methods to do both phases in one go.
Avoids confusing overload behavior when constructing QgsRasterLayers
from c++ code. This extra constructor was causing character literals
to be converted to a boolean loadDefaultStyle flag instead
of being used as the provider key. Removing the extra constructor
avoids this, and removes some duplicate code.
... rather than using QgsProject::instance() internally
These are small cleanups to dig out some instance() uses and move them one level up...
At some point we should maybe make labeling engine configuration a part of QgsMapSettings
and have default project labeling engine config accessible from QgsProject
in a way similar to e.g. snapping configuration.
