This patch will make Mapbox GL Core never use OpenGL
directly. We should consider locking into OpenGL ES 2.0
to simplify the code path and remove #ifdefs.
Commit 1c53dc9ca8 ([core] Refactor vertex array object extension
initialization) changed where the GL renderer string is checked in order
to prevent the vertex array extension initialization in certain
platforms and in doing so made the logging of that string more frequent,
once per context initialization instead of at most once per process.
This is a lot more verbose, specially when running unit tests, so make
sure that string is logged at most once as it was previously.
The ANGLE implementation of GL_KHR_debug on Direct3D seems to be broken
leading to a bad_alloc being thrown when running with a debug build, so
disable it in that platform.
Currently the vertex object extension is disabled through an ifdef for
the Windows platform due to an issue with ANGLE, while there is a
blacklist for other platforms. Unify those by adding ANGLE to that
blacklist and some small refactoring.
Add a parameter to Program::draw to control whether face culling should
be enabled. This will be used in a follow up commit to enable face
culling for fill extrusion layers.
Port of https://github.com/mapbox/mapbox-gl-js/pull/5101: adds a new render pass `Pass3D` before any other rendering wherein we render layers with 3D passes (fill-extrusion layers) to offscreen framebuffers, sharing a depth renderbuffer between those layers in order to render 3D space correctly. Those framebuffers are saved on the RenderLayers and copied back to the map during the translucent pass. Rendering to offscreen framebuffers before we do any clear + draw means we can avoid expensive framebuffer restores.
When a texture is deleted, it is as if all texture units to which that texture object is bound are rebound to texture object zero. Therefore it's the *bindings* that need to be dirtied, not the active texture unit.
These changes are necessary for programs whose set of active attributes is not fixed at compile time by a template parameter pack, but rather varies based on the generated shader text at runtime. In such cases, the attribute location of a given named attribute may vary between instances of the same Program.
Previously, attribute bindings were implicitly associated with a location based on template parameter order, and -1 was used to indicate an inactive attribute. This left us unable to disable the appropriate attribute when it went from active to inactive. Now, the state tracker for bindings explicitly associates locations and state, and an empty optional is used to indicate an inactive attribute.
In addition, a gl::VertexArray class is now exposed, allowing more flexibility in the relationship between Programs, Segments, and attribute bindings. In this commit, that relationship does not change, but the subsequent commit adjusts it to match gl-js, reduce rebinds, and work around buggy VAO implementations.
VertexArray uses a pimpl idiom in order to support implementations that lack the VAO extension. In that case, all VertexArrays share global binding state, reflecting the platform reality in the absence of VAOs, while still providing a uniform API.
port https://github.com/mapbox/mapbox-gl-js/pull/4781
This improves legibility of labels that follow lines in pitched views.
The previous approach used the limited information in the shader to
calculate put the glyph in approximatelyright place. The new approach
does this more accurately by doing it on the cpu where we have access to
the entire line geometry.
Previously, we initialized global variables that held pointers to the extension functions. While this seemed to work, the spec doesn't guarantee that the function pointers are identical for different OpenGL contexts. Therefore, we are now making them a member variable of the Context object.
This is safer now -- it can be an implementation detail of Segment/Context. And the typing of SegmentVector now ensures that the attributes and program match.
