[PATCH] D27486: Correct class-template deprecation behavior
Aaron Ballman via Phabricator via cfe-commits
cfe-commits at lists.llvm.org
Wed Jan 4 09:40:39 PST 2017
aaron.ballman added inline comments.
================
Comment at: include/clang/Basic/Attr.td:301
bit DuplicatesAllowedWhileMerging = 0;
+ // Set to true if this attribute should apply to template declaration,
+ // remains false if this should only be applied on definition.
----------------
template declaration -> template declarations
================
Comment at: include/clang/Basic/Attr.td:302
+ // Set to true if this attribute should apply to template declaration,
+ // remains false if this should only be applied on definition.
+ bit AppliesToTemplateDeclaration = 0;
----------------
on definition -> to the definition
================
Comment at: include/clang/Sema/Sema.h:7388-7391
+ void InstantiateAttrsForDecl(const MultiLevelTemplateArgumentList &TemplateArgs,
+ const Decl *Pattern, Decl *Inst,
+ LateInstantiatedAttrVec *LateAttrs = nullptr,
+ LocalInstantiationScope *OuterMostScope = nullptr);
----------------
Did clang-format produce this formatting?
================
Comment at: lib/Sema/SemaTemplateInstantiateDecl.cpp:320
+ CXXRecordDecl *ThisContext =
+ dyn_cast_or_null<CXXRecordDecl>(ND->getDeclContext());
+ CXXThisScopeRAII ThisScope(*this, ThisContext, /*TypeQuals*/ 0,
----------------
Are you sure `ND` is always non-null? If so, then you should use `cast<>` above instead of `dyn_cast<>`.
================
Comment at: utils/TableGen/ClangAttrEmitter.cpp:2456
+void EmitClangAttrTemplateInstantiateHelper(const std::vector<Record *> &Attrs,
+ raw_ostream &OS, bool DeclTime) {
+ OS << " switch (At->getKind()) {\n";
----------------
What does "time" mean in `DeclTime`?
================
Comment at: utils/TableGen/ClangAttrEmitter.cpp:2507
+}
+// Emits code to instantiate dependent attributes on templates.
+void EmitClangAttrTemplateInstantiate(RecordKeeper &Records, raw_ostream &OS) {
----------------
Add a newline above this comment to separate it from the previous function body.
https://reviews.llvm.org/D27486
More information about the cfe-commits
mailing list