[PATCH] D90763: Traverse-ignore explicit template instantiations

Aaron Ballman via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Thu Nov 5 12:54:24 PST 2020


aaron.ballman added inline comments.


================
Comment at: clang/lib/ASTMatchers/ASTMatchFinder.cpp:503
                           TraversalKind Traversal, BindKind Bind) {
+    auto ScopedTraversal = TraversingTemplateInstantiationNotSpelledInSource;
+
----------------
Please spell this type out rather than use auto.


================
Comment at: clang/lib/ASTMatchers/ASTMatchFinder.cpp:506
+    if (const auto *CTSD = Node.get<ClassTemplateSpecializationDecl>()) {
+      auto SK = CTSD->getSpecializationKind();
+      if (SK == TSK_ExplicitInstantiationDeclaration ||
----------------
Same here, though this could also be simplified to:
```
ScopedTraversal = (SK == TSK_ExplicitInstantiationDeclaration || SK == TSK_ExplicitInstantiationDefinition);
```


================
Comment at: clang/unittests/AST/ASTTraverserTest.cpp:1092
+
+// Explicit instantiation of template functions do not appear in the AST
+template float timesTwo(float);
----------------
Huh, do you have any idea if that's a bug? We have `ClassTemplateSpecializationDecl` and `VarTemplateSpecializationDecl`, but we have `FunctionTemplateSpecializationInfo` that doesn't generate an AST node and no mention of why in the comments that I've spotted yet.


================
Comment at: clang/unittests/ASTMatchers/ASTMatchersTraversalTest.cpp:2280
+    EXPECT_TRUE(matches(Code, traverse(TK_AsIs, M)));
+    EXPECT_TRUE(matches(Code, traverse(TK_IgnoreUnlessSpelledInSource, M)));
+  }
----------------
Explicitly instantiating a function template in ignore mode returns false, but explicitly instantiating a class template returns true? Is this intentional or just fallout from the lack of explicit instantiation information in the AST for functions?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D90763/new/

https://reviews.llvm.org/D90763



More information about the cfe-commits mailing list