GBitmap Changes:
Added all other formats to gbitmap that we support
gbitmap now supports cubemaps
added converters for all these other formats
added stb_image_resize for extrudemips so we can extrude mipmaps for all other formats
GFXTextureManager
Can now directly make cubemaps and texture arrays based on the GFXTextureProfile
API implementations for all functions that cubemaps and arrays needed
asset brower previews were asserting due to a cornercase rt assignment
also ditch mipgen for rts targetting a texture profile that's marked as nomip, and adjust hdr uopsampling shader to compensate for overdarks
dx11 has a 2048/6 cap on how many cubemaps can fit in a given array. bump the probe count up to just under that to buy folks head room while we work on revising a more robust solution
the key part of this issue was to invert the order we were baking the faces. Not much of a fix but it works better than before need to get a proper fix implement
the way the issue is copying X+ into every other face but isnt doing it in the reverse order means the rtv for face 0 must be being replicated into all other faces.
now shaders ask the device to create a buffer, if a buffer of the same name and size already exist then a pointer to that buffer is returned instead. Going to see if something similar can be added to the opengl end.
opengl can now compile with ubo buffer objects similar to cbuffers on dx side.
cleaned double up of data from both sides, gfxhandles only need to use the desc info instead of holding onto its own.
Reverted back to a 2d array since when the const buffer changes are made to the shaders they will be completely swapped out for a 1d array with a max size of 16 for the program, id3d11buffers will be shared across shaders apart from object buffers. We still need to decide on a structure to get everything setup so to keep functionality a max of 2 buffers per stage with 1 buffer as a padding so we can start coding constant buffers right away
need to reverse the bitshift on shaderstages so we get values between 0-5, need to also revert the mBoundConstVec back to a 2dimensional array so that the binding point can be dynamic as well.
Added multiple descriptions to handles.
While maps are very efficient and match data very quickly due to hashing and what not, they are tricky memory wise, we cant clear them if they are only made up of structs and enums. Need to explore other methods for doing this. Original system had a sub buffer system with a check based on whether the handle was marked vertex/pixel. With 6 stages possible now we need to think of a more effective method. Since shaderstage is now a bit flag maybe some sort of streaming setup could be used to write without a loop.
working example, renders all gui controls.
We still cant render a scene, something to do with the constant buffer at slot 0 of the vertex shader being too small..... gotta figure this one out..................
Refactored DX Shader compilation and reflection to be more similar to opengl
Added support for more than 1 const buffer per shader stage.
This is boilerplate code, no functionality yet, that will be added in further commits. Though this builds without errors do not try and run it at this point.
The fix adds a sanity check to the D3D and GL device enumeration function that exits early if there's no registered display from the OS(implying it's running in commandline).
With newer versions of SDL, attempting to enumerate the GL device causes a segfault because the OS will return back a valid context, but the context isn't actually valid. So when tested against, it crashes. Avoiding enumerating the device when you're not in a position to render works around the issue.
TODO: If the machine is running in terminal-only mode, and thus has no valid render context, but still has a monitor plugged into the machine, it can sidestep the sanity check and still result in a segfault. Need a more robust check for that circumstance in a future fix.
-Adds reversed depth projection model, dramatically increasing depth buffer effective resolution.
-Adds 32F depth 8U stencil format GFXFormatD32FS8X24 (following DX naming conventions). Note this is a 64-bit format, and likely not suitable for mobile platforms. Revert to GFXFormatD24S8 in renderManager.tscript for mobile & "ancient" platforms.
-Corrects alignment of texture type details array.
