Positions were originally done using Point2D/3D to simplify their
transformation using matrices and to some extent simplify their usage in
shaders. But now the disadvantages exceeded the advantages:
* They take 50% more for 2D positions and 33% more for 3D positions, as
last coordinate is always equal to 1, on the other hand when last
coordinate is errorneously not equal to 1 they have crazy behavior.
* Normalizing them or transforming them with anything else than with
matrices is PITA, as we need to strip the last component, do the
transformation, and then add the component back.
* All transformation handling classes (Complex, DualComplex,
Quaternion, DualQuaternion, Matrix3, Matrix4) now have convenience
functions for transforming points specified directly as
Vector2/Vector3 (and also for transforming vectors).
* When someone wants to use homogeneous coordinates with crazy last
component values, they can do so with plain Vector3 for 2D and
Vector4 for 3D and it will be less confusing than using Point2D/3D
which no important detail hidden.
Removed workarounds for alias templates, variadic templates and
anonymous enums, but 1.8.2 has some bug with forward declarations
causing classes to appear in default namespace, breaking
cross-references.
* Calling *Mesh::draw() with parameter to start on arbitrary vertex
index might give users more freedom than they want to have (e.g.
lines rendered where gaps should be, broken triangle strips...).
* Single-precision floats have meaningful precision of ~6 decimal
places, everything after that would be random garbage anyway, so we
don't need anything "more precise" for icosphere.
* Texture1D can have only one target and it can be used as framebuffer
target.
It prevents unwanted implicit conversions from e.g. nullptr to Camera,
Vector2 to Physics::Point etc. By making all the constructors explicit
it is easier to routinely add the keyword to all new classes instead of
thinking about cases when to add and when not to.
Magnum.h now doesn't include anything except OpenGL headers, thus
changes in Math library don't trigger recompilation of everything, but
only of things really depending on it.
Math constants moved to separate file for similar reasons, de-inlined
some functions to remove the need for some #includes.
It only caused another maintenance burden and was confusing to users.
Now when scene graph is in SceneGraph namespace there is no need for
another grouping. Namespaces are (and should be) sufficient.
This reverts commit 79945ab6fc.
Conflicts:
src/BufferedImage.h
src/BufferedTexture.h
src/Framebuffer.h
src/Query.h
src/SceneGraph/Scene.h
While namespaces act for hierarchy, modules are something like "tags" -
usable when you want to check related classes of e.g. CubeMapTexture.
Not sure how to name module for Math and Physics namespaces and
Contexts/Trade, though.
* Renamed Type enum to Target to be consistent with OpenGL naming
* Strongly typed enums
* Ability to specify different type when binding/setting data
* Added missing buffer targets
* Removed VertexPointer typedef, as size_t is somewhere 64bit and
OpenGL doesn't have corresponding data type. Using unsigned int
everywhere instead, to avoid confusion.
* Removed Face structure altogether, using vertex indices directly.
* Greatly reduced size of unit tests, also thanks to C++11
std::initializer_list feature.
Vladimír Vondruš