Websites built with tons of polyfills are likely not run on these devices anyway, since they will run out of RAM before, let alone after they will only load after sone minutes because of CPU limitations on top of not being loaded because their x509 certs are outdated as well as the bandwith they support is not suitable to load MB sited pages
Actually the author is building onto my suggestion https://github.com/soasis/idk/pull/17/files which I created several years ago, but adapted a bit too fast while submitting. And yes it looks very similar to GNU nested functions, since I started tinkering with these first.
I am really not sure if all these observations mentioned in the article are 100% correct, though.
First : Code seems to be compiled with clang. On Linux with gcc the native function one is way faster than the clang one.
Second: The author does run the code on ARM64/MacOS .
At least on my ryzen CPU on Linux with gcc the "normal C code" is way faster than anything else. Not that we do not need to thing about "closure" type functionality, but one should be careful to extrapolate implementations from one compiler on one platform to the rest of the pack.
Regarding N3654 I am not sure how to benchmark it here, since C could potentially use __builtin_call_with_static_chain , but I am not sure how to write the function to use the chain for accessing the variables.
I tried to estimate N3654 it by using "tinygo" which is AFAIK using the usual Calling ABI, but it was a factor of two slower than clang. Even "go" with its very specific ABI is still much slower. I discovered this isn't representative since runtime calling costs had been totally shadowed by costs of allocations.
I had the chance to test Moment while my Widex Evoke 440 had been in repair. They are indeed a lot better as the Evoke, sounding wise, while having only mild hearing loss. But they are still only MFI (can only connect to an iphone) and since I am migrating away from Apples ecosystem this will be a blocker.