This commit adds new functionality to NativeScript, namely:
- ability to set and get documentation for classes, methods,
signals and properties
- ability to set names and type information to method arguments
- ability to set and get type tags for nativescripts
- ability to register instance binding data management functions
- ability to use instance binding data
The GDNative C API gets passed to libraries in a struct of function
pointers. To provide stable binary compatibility, each extension not
part of the core API is separated into its own sub-struct.
These structs aren't meant to be changed in order to keep binary
compatibility.
In case of an API extension, the structs include a `next` pointer
which can point to a new struct with additional function pointers.
Godot's build system generates the API structs automatically at
build time, but so far there has no support for the mentioned `next`
pointers.
This commit changes the API struct generation in such a way that code
that used previous headers will compile without problem with the new
headers.
The new extension-extensions (weird name, but that's what it is) get
generated recursively and include the version in the struct-name.
The GLES3 shader compiler performs certain checks to enable or disable
the usage of certain uniform variables (and with that the set-up of UBOs).
If the `TIME` variable gets used inside the `vertex` function then the
renderer knows that it has to insert that value into the UBO.
The same applies to the `fragment` function.
The `light` function gets executed inside the fragment shader for every
light source that is relevant to the current pixel. If the `TIME` variable
gets used in that function then it needs to be present in the fragment-UBO.
The check for this was missing, so if a shader uses `TIME` inside `light`
but not inside `fragment` then the uniform will not actually be set up.
The ear clipping algorithm used to triangulate polygons has a slightly too conservative point-in-triangle test which can, in some configurations prevent it from finding a possible tessellation. Relaxing the test by considering that points exactly on edges don't belong the triangle fixes the issue. Changing the semantic of the test is safe because no other code makes use of it. A more detailed explanation can be found in issue #16395.
Fixes#16395.
When running the engine with -d we get a message on the command-line for
each control being clicked. After discussing with @reduz it seems that
this is old and should be removed. Commented out as requested.