erichkeane wrote: > > Requesting changes just to be sure none of the changes from `createRef()` > > to `createDup()` would have generated a `StringRef` holding a null pointer. > > I'd be surprised if they did, but a second set of eyes would be helpful. > > I am not sure how to proceed here. This change was requested by > @efriedma-quic on #125020 > > I see 3 options: > > 1. Land as-is, but then we would not differentiate `createNull()` vs > `createEmpty()` vs `createDup()`. No idea if this difference is important. > > 2. Move checks into `createDup()`, but then `createDup()` is not always a > "Dup". Also no idead if this is important. > > 3. Just abandon the patch. > > > WDYT?
Aaron is away this week, but: I believe he was asking for an analysis of all of the uses to make sure they don't use a null-StringRef anywhere. IMO, an assert of some sort in the createDup interface would be sufficient to long-term prevent the problem. https://github.com/llvm/llvm-project/pull/126683 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
