[PATCH] D82278: Fix traversal over CXXConstructExpr in Syntactic mode

[Aaron Ballman via Phabricator via cfe-commits]

aaron.ballman added reviewers: rsmith, sammccall.
aaron.ballman added inline comments.


================
Comment at: clang-tools-extra/clang-tidy/bugprone/UnusedRaiiCheck.cpp:80
   auto Matches =
-      match(expr(hasDescendant(typeLoc().bind("t"))), *E, *Result.Context);
+      match(traverse(TK_AsIs, expr(hasDescendant(typeLoc().bind("t")))), *E,
+            *Result.Context);
----------------
While I believe the change is necessary here, it's not obvious to me what "hints" tell me this behavior is correct for the given matchers. None of the matchers look like they're going to care about implicit nodes, so how am I to know that AsIs is correct or not as a reviewer? As it stands, I sort of feel like I have to take it on faith that this change is correct and it looks a little suspicious because the code using the matcher is creating a fix-it at what now may be the location of an implicit node.


================
Comment at: clang/lib/AST/Expr.cpp:3001
         Expr *A = C->getArg(0);
-        if (A->getSourceRange() == SR || !isa<CXXTemporaryObjectExpr>(C))
+        if (A->getSourceRange() == SR || C->isElidable()) {
           E = A;
----------------
Looks like the change introduced new curly braces for a single-line if.


================
Comment at: clang/lib/AST/ParentMapContext.cpp:157
 
+      if (auto *C = dyn_cast<CXXFunctionalCastExpr>(E)) {
+        if (C->getSourceRange() == SR)
----------------
`const auto *`


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D82278/new/

https://reviews.llvm.org/D82278