rmsyn
> hence if you limit to "if your system can run..." then the MSRV can absolutely be older Fair enough, but you see the contradiction here? The MSRV to get...
I've done the work in my fork, if you want to crib some stuff:
> not sure on an ETA for a alpha release of the crate but this should help both support users who don't want to panic at all and those who...
@repnop I added some changes to address the points you brought up. I removed the `Option` wrapper from the return type. Any failure to produce `iter` items will now just...
## `jalr` instead of `jal` This has been resolved since this issue was opened: ## `noreturn` `nounwind` and divergent functions This has been resolved since this issue was opened: ##...
> The compiler fence seems to disappear when -C opt-level="..." is passed. Confirmed that this works, `atomic::compiler_fence` disappears, where `atomic::fence` emits a `fence` instruction. So, I guess this would only...
If this is approved, I can open another PR to use these selectors in the rest of the library, or add more commits to do it here.
> Regarding using target_arch instead of custom compilation flags, I don't know what to say. How are these custom compilation flags added currently? I ran into this issue because of...
Added some fixes for the rest of the `riscv` crate. Let me know if you would like me to split these changes into a separate PR instead.
> However, changing the build script as suggested by @taiki-e sounds more than reasonable. I was exploring this a bit, as well. I'll try the suggestion by @taiki-e. If it...