There are some issues currently preventing Miri runs.
While attempting to fix those and run miri, I found other problems in various code paths, Will use this ticket as a centralized board for them.
Duplicate String Writing:
serde::se::pp::test::prettyfy failed because it wrote duplicate strings due to a bug in value-trait:
Issue has a PR attached as well
simd-lite/value-trait#62
Calling alloc(0) is undefined behaviour and should be forbidden and handled.
#442
Calling .as_mut_ptr() numerous time on a slice is not allowed as Miri's borrow checker considers the mut borrow invalidated:
#443
Unrolling using SSSE3 features does not require SSSE3 at runtime, and may reach "invalid instruction":
#444
No UTF-8 validation happening at all in one of the code paths:
#445
After these, the only thing blocking from running miri is a couple of things marked unused for some reason when running miri, did not debug too deeply, but just adding #[cfg(not(miri))] on them was enough to get it to run
There are some issues currently preventing Miri runs.
While attempting to fix those and run miri, I found other problems in various code paths, Will use this ticket as a centralized board for them.
Duplicate String Writing:
serde::se::pp::test::prettyfy failed because it wrote duplicate strings due to a bug in
value-trait:Issue has a PR attached as well
simd-lite/value-trait#62
Calling alloc(0) is undefined behaviour and should be forbidden and handled.
#442
Calling .as_mut_ptr() numerous time on a slice is not allowed as Miri's borrow checker considers the mut borrow invalidated:
#443
Unrolling using SSSE3 features does not require SSSE3 at runtime, and may reach "invalid instruction":
#444
No UTF-8 validation happening at all in one of the code paths:
#445
After these, the only thing blocking from running miri is a couple of things marked unused for some reason when running miri, did not debug too deeply, but just adding #[cfg(not(miri))] on them was enough to get it to run