this is brilliant. iOS Deployment Target is the operating system level. if you add a new extension to an old app, it will have a deployment target (OS level) set to the highest that the Xcode version that you are using supports, by default. however, if your test device is at a lower OS level, the extension will not show and you won't see any errors reported. silent fail. this answer fixed it for me.

app store connect binary upload failing again, same error mar 28 2022

thanks. i see some notes online that there is only a private SDK for this neural engine, nothing we can use as external devs. such an enticing name but no access. here is tasty donut, but you can't have it.

thanks, will have to get back to this in the near future, we had to put it aside because the signing looks to be non-trivial. it looks like the method we are using to build the underlying code ("pyinstaller" ... python installer) packs a whole bunch of .so and .dylib libraries together in a directory hierarchy, and each one of these need to be signed, possibly. yikes.

ah ha, thanks

hello, thanks for the response! no i dont need drand48() particularly, that is just my go-to on apple platforms. any reasonable uniform random [0.0, 1.0] would work. i saw some other ones in the docs but havent tested them for this thread interference issue. one thing about the random functions that i noticed, they can get slower if they are "truly" random. i just need approximately random.

sorry i have not been getting emails from these forums here you go! https://github.com/diffent/videocolorissue