[compiler-rt] [llvm] [msan] Automatically print shadow for failing outlined checks (PR #145107)

[Vitaly Buka via llvm-commits]

vitalybuka wrote:

> > Why do we want to expose out internals into UI? I don't believe it's useful to end-users, BTW same for Asan shadow, but it's already there.
> 
> I believe having it simplifies debugging - it saves users the need to call `__msan_print_shadow` (which is documented externally at https://github.com/google/sanitizers/wiki/memorysanitizer#interface and internally in Big Tech's MSan landing page). Seeing whether the shadow is entirely uninitialized or partly initialized, or the pattern of partial initialization, can often give hints as to where the missing initialization is.

It's going to show up in very limited cases, even if it shows up it will affect the user decision only a tiny fraction of those cases. 

In cases where the user can now understand the reason easily now, after the change they will guess what this is for.
So it's likely regression on average.



https://github.com/llvm/llvm-project/pull/145107