jbcoe added a comment.
I think there are some minor issues with the tests, easily fixed. I wonder if
the matchers can catch things inside implementation-defined inline namespaces
like `std::_v1_::experimental::library_fundamentals_v1::string_view`.
================
Comment at: clang-tidy/misc/DanglingHandleCheck.cpp:57
@@ +56,3 @@
+ast_matchers::internal::Matcher<RecordDecl> isASequence() {
+ return hasAnyName("::std::deque", "::std::forward_list", "::std::list",
+ "::std::vector");
----------------
Will this (and similar checkers) handle inline namespaces?
================
Comment at: test/clang-tidy/misc-dangling-handle.cpp:73
@@ +72,3 @@
+ StringRef(const char*); // NOLINT
+ StringRef(const std::string&); // NOLINT
+};
----------------
What does `NOLINT` do here?
================
Comment at: test/clang-tidy/misc-dangling-handle.cpp:85
@@ +84,3 @@
+ std::string_view view_2 = ReturnsAString();
+ // CHECK-MESSAGES: [[@LINE-1]]:29: warning: std::basic_string_view outlives
+
----------------
This (and other) check-messages line looks truncated.
http://reviews.llvm.org/D17811
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits