So far not reproducible on newer xcode versions.
cc @varkor @kennytm
52dec0e obviously did not fix it and can be reverted.
Happens inside
rustc::ich::hcx::_$LT$impl$u20$rustc_data_structures..stable_hasher..HashStable$LT$rustc..ich..hcx..StableHashingContext$LT$$u27$a$GT$$GT$$u20$for$u20$syntax_pos..span_encoding..Span$GT$::hash_stable::h069c687f17acced0 + 760
whenever it happens, but it's random during which crate's compilation it happens or whether it occurs at all.
Guesses so far are
- alignment issues with
Span, which is repr(packed) and only has a u32 field
- something to do with the
CACHE thread_local (just because miri touched statics)
Is is possible to produce better stacktraces (with line info e.g.) on macs?
So far not reproducible on newer xcode versions.
cc @varkor @kennytm
52dec0e obviously did not fix it and can be reverted.
Happens inside
whenever it happens, but it's random during which crate's compilation it happens or whether it occurs at all.
Guesses so far are
Span, which isrepr(packed)and only has au32fieldCACHEthread_local (just because miri touched statics)Is is possible to produce better stacktraces (with line info e.g.) on macs?