[PATCH] D84600: [Analyzer] Support note tags for smart ptr checker

[Nithin VR via Phabricator via cfe-commits]

vrnithinkumar added inline comments.


================
Comment at: clang/lib/StaticAnalyzer/Checkers/SmartPtrChecker.cpp:110
+  SmartPtrChecker *Checker = Mgr.registerChecker<SmartPtrChecker>();
+  NullDereferenceBugTypePtr = &Checker->NullDereferenceBugType;
 }
----------------
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`


================
Comment at: clang/lib/StaticAnalyzer/Checkers/SmartPtrChecker.cpp:95-96
+
+  const auto *RecordDecl = MethodDecl->getParent();
+  if (RecordDecl && RecordDecl->getDeclName().isIdentifier()) {
+    OS << " of type '" << RecordDecl->getQualifiedNameAsString() << "'";
----------------
NoQ wrote:
> vrnithinkumar wrote:
> > NoQ wrote:
> > > Aha, this time you checked for anonymous declarations! Good.
> > > 
> > > That said, i'm not sure type is actually useful for this warning, because they're roughly all the same.
> > Okay. I was trying to be consistent with `MoveChecker`.
> > Just removed smart pointer type.
> These were important in `MoveChecker` because use-after-move rules for smart pointers are different from use-after-move rules of any other class in the standard library and the checker has to behave differently (https://wiki.sei.cmu.edu/confluence/x/O3s-BQ).
Oh
That makes sense.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D84600/new/

https://reviews.llvm.org/D84600