voyager-jhk wrote: > It seems like I had some ideas about the linked issue and what it could take > to fix it. This PR looks significantly different to what I pitched there and > that makes me wonder how these two approaches relate to each other and why > did you settle on this one. This wasn't apparent on the issue thread not in > the PR summary here.
I kept the change in `MallocChecker` because changing the general escape-on-bind behavior would have a wider impact across checkers, and I wasn't sure how to make that change safely within the scope of this issue. The current fix preserves tracking only when the parent allocation is still tracked as allocated. This seemed like a more contained way to address the false negative, while leaving the broader escape behavior unchanged. https://github.com/llvm/llvm-project/pull/221587 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
