NoQ added inline comments.
================ Comment at: clang/lib/StaticAnalyzer/Checkers/SmartPtrChecker.cpp:110 + SmartPtrChecker *Checker = Mgr.registerChecker<SmartPtrChecker>(); + NullDereferenceBugTypePtr = &Checker->NullDereferenceBugType; } ---------------- vrnithinkumar wrote: > NoQ wrote: > > Wait, i don't understand again. You're taking the bug type from that > > checker and using it in that checker. Why did you need to make it global? I > > thought your problem was about capturing the bug type from the //raw// null > > dereference checker? > > Wait, i don't understand again. You're taking the bug type from that > > checker and using it in that checker. Why did you need to make it global? I > > thought your problem was about capturing the bug type from the //raw// null > > dereference checker? > > The check for bug type is in `SmartPtrModeling` but the bug type is defined > in `SmartPtrChecker` Aha, ok, i misunderstood again. So i guess we didn't need a new header then. That said, it's an interesting layering violation that we encounter in this design: it used to be up to the checker to attach a visitor and so should be activating a note tag, but note tags are part of modeling, not checking. I guess that's just how it is and it's the responsibility of the checkers to inform the modeling about bug types. I guess the ultimate API could look like `BugReport->activateNoteTag(NoteTagTag TT)` (where `TT` is a crying smiley that cries about "tag tags" T_T). But that's a story for another time. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D84600/new/ https://reviews.llvm.org/D84600 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
