steakhal added a comment. In D125709#3518242 <https://reviews.llvm.org/D125709#3518242>, @martong wrote:
>> I'm not sure, shall I add tests? > > Yes, please. Unit tests for `dyn_cast` and `isa` should be easy. However, I > am not sure how to test `cast` for the failure cases. What if we were using the 'new' casting style at some places. If it compiles, it should be fine. It feels weird to check stuff, only in `static_asserts`. In D125709#3519033 <https://reviews.llvm.org/D125709#3519033>, @bzcheeseman wrote: > In D125709#3518096 <https://reviews.llvm.org/D125709#3518096>, @steakhal > wrote: > >> I had to fix the `doCast` to return `To` instead of `Optional<To>` to make >> it work. > > That's fine (or it should be!), you could dereference the optional if you > wanted to Currently, we expect that casts result in regular SVal objects, instead of pointer-like objects , thus this code to compile: `NonLoc N = llvm::cast<NonLoc>(V)`, where `V` is of type `SVal`. I believe that is why I decided to make that change. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D125709/new/ https://reviews.llvm.org/D125709 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
