[llvm] [SDAG] Introduce inbounds flag for ISD::PTRADD (PR #162477)

[Eli Friedman via llvm-commits]

efriedma-quic wrote:

> and the inbounds-ness of the outer ptradd / the last gep hinges on the something_... addition that was optimized away?

Slightly more realistically, you'd have pointers to objects that are adjacent, or have non-overlapping lifetimes, that happen to compare equal, and you swap one integer value for the other.  But that's the basic idea, yes.

https://github.com/llvm/llvm-project/pull/162477