[llvm-dev] Should analyses be able to hold AssertingVH to IR? (related to PR28400)

[Sanjoy Das via llvm-dev]

Hi Sean,

Sean Silva wrote:
 >
 >     But asan won't catch problems (insofar I understand how it works) if
 >     the free'ed BasicBlock is used as a key in a DenseMap or something --
 >     if another BasicBlock gets allocated to the same location we'll 
end up
 >     returning bad cached results.
 >
 >
 > ASan's allocator is specifically hardened against reusing memory, which
 > mitigates this, but I'm not sure by how much (but I think by a
 > significant amount).

Do you mean it will re-use memory less often?  Won't that just hide
the bug above?  If anything, I don't want ASan to "mitigate" bugs, I
want it to make the bug trigger more often. :)

(Or did I just re-state what you were saying?)

-- Sanjoy