This fixes a conflicting definition of GLsizeiptr and GLintptr when
Magnum is combined with qopengl.h from Qt on Windows. Khronos has it as
`signed long int`, whereas qopengl.h as ptrdiff_t, which is just `int`.
This one explicitly loads GL 1.0 and 1.0 function pointers on EGL
contexts on NVidia drivers (Linux headless boxes), because somehow the
usual statically linked functions don't behave correctly.
Minimal updates (just the include guards) so Git is hopefully able to
detect the rename and track the history properly.
Everything except Magnum::GL doesn't compile now.
And removing the bundled std::optional implementation. This finally
makes this library compatible with C++17. Since this would be a huge
backwards-incompatible change that would make everyone angry, the
following had to be done in case both CORRADE_BUILD_DEPRECATED and
MAGNUM_BUILD_DEPRECATED is defined:
* Under C++11 and C++14, Containers::Optional / Containers::NullOpt is
aliased to std::optional / std::nullopt. This is no worse than the
state before, when we also provided these symbols.
* Under C++17, where standard <optional> header is available,
Containers::Optional provides implicit conversion to it. Only one-way
conversion is supported, as there was fortunately no Magnum API that
took std::optional via parameter, and there might be some corner
cases that this doesn't cover. The goal is to have all examples
compiling with the old API, at least.
* There's a new test especially for this, which checks that both the
C++11 and C++17 ways of doing things work as they should.
The typedef and conversion is marked as deprecated, so it will spit out
many warnings to push users to upgrade. I hope I can completely remove
this mess soon :/
Because the library still links to the old crappy opengl32.dll, we need
to load all symbols above OpenGL 1.1, not just those that are above
OpenGL ES 2.0/3.0.
As with Corrade, this is not exactly backwards compatible, but for
common use case without OBJECT libraries this should not be a problem.
In any case, recreate the build dir and update your copy of all
Find*.cmake modules to avoid weird things happening.
User-facing changes:
* Documentation of all Find*.cmake modules converted to
reStructuredText to follow official CMake guidelines.
* The newfangled way to use the libraries is to link to Magnum::Shaders
instead of adding ${MAGNUM_SHADERS_INCLUDE_DIRS} to include path and
linking to ${MAGNUM_SHADERS_LIBRARIES}.
* The old ${MAGNUM_*_LIBRARIES} are deprecated and now just expand to
Magnum::* target. Use the target directly. These are also enabled
only when building with MAGNUM_BUILD_DEPRECATED.
* The old ${MAGNUM_*_INCLUDE_DIRS} are removed as the Magnum::* targets
cover these too.
Internal changes:
* Global state such as include_directories() was replaced with
target-specific settings.
It seems that there is no such thing as eglGetProcAddress() and thus we
are able to use only functions that are defined in the ES[23]/glext.h
header and no others. So, currently, the function loader internally
undefs all function name macros that were defined in our flextgl.h
header, then includes the Apple's glext.h header and assigns function
pointers of those extensions that are defined in the header. Apple also
has some minor differences in function signatures (different constness
of pointer-to-pointer variables) so I had to reinterpret_cast
everything.
Might seem to be uglier than including glext.h directly in our code,
but I made bad experience when doing so -- I want to depend on my
header bugs that are consistent across all platforms instead of
depending on whatever changes Apple makes in its headers. I also want to
have all functions defined and not only those that are supported on
iOS.
Phew. Thank you, flextGL, for making it way easier than it appeared to
be at first.