split shader files across multiple QRC files, partially fix 4.1

[HifiExperiments]

fixes #2044 + (partly) #1570

[ada-tv]

Code diffs look alright to me, the 410 renderer works fine now (tested with MESA_GL_VERSION_OVERRIDE=4.1)

Saw the handful of split RCC calls running while compiling

450, es310, and vk all work fine too

only tested on linux, leaving the QA tag on until someone tests on windows too, though i'm confident it'll work there too

[HifiExperiments]

hrm actually I'm now seeing something weird in the deferred renderer (on windows) were most shaders are just rendering black (but not all? most unlit shaders seem fine). hadn't noticed because forward renderer works. not sure if it's because of this PR or something that went in recently (vulkan?). @ada-tv does deferred renderer work for you with this on linux? maybe will try a renderdoc capture when I get a chance

[ada-tv]

hmm yeah, deferred is all black

forward	deferred

the albedo and normal buffers look ok, but the result of RenderDeferredSetup is just black

inputs	outputs

there's a lot of "x used uninitialised" warnings from the shader compiler, tho the identifiers have been stripped from the optimised glsl

[HifiExperiments]

@ada-tv re-disabling the SPIRV shaders in Shaders.cpp seems to have fixed it. something must have broken about the SPIRV optimization between #1531 and now. but we can fix that separately - I think this should be good to go now

[ada-tv]

Forward and deferred work properly now on 450, es310, and vk.

410 doesn't work though, likely an unrelated regression related to the new layout macros the Vulkan PR introduced: unrecognized layout identifier `binding'

[ksuprynowicz]

I tested it on Windows with OpenVR and OpenXR and everything works well on OpenGL 4.5

[HifiExperiments]

Forward and deferred work properly now on 450, es310, and vk.

410 doesn't work though, likely an unrelated regression related to the new layout macros the Vulkan PR introduced: unrecognized layout identifier `binding'

ah yeah binding is a 4.2 feature. we might just need to remove the changes to the 4.1 header from da01b4c

[ksuprynowicz]

@ada-tv @HifiExperiments Is this PR safe to merge in your opinion?

[HifiExperiments]

yeah - it fixes building on my machine, and I’ve updated the title to reflect that it only partially fixes 4.1. I can look into the remaining issue with 4.1 separately

[ksuprynowicz]

Thank you!