> Guillaume, I decided not to include the test you suggested since it fails,
but the reason is not the the inline problem because if I don't use macros it
works just fine, there's a test in iterator.cpp for std::vector. I'll log
another bug for that.
I don't see `USE_INLINE_NAMESPACE=1` in //iterator.cpp//, so I don't think it
tests the bug.
What about this test?
// RUN: grep -Ev "// *[A-Z-]+:" %s > %t.cpp
// RUN: cpp11-migrate -use-auto %t.cpp -- --std=c++11 -I %S/Inputs
// RUN: FileCheck -input-file=%t.cpp %s
#define USE_INLINE_NAMESPACE 1
#define CONTAINER vector
#define USE_INNER_CLASS_ITERATORS 1
#include "test_std_container.h"
#undef USE_INNER_CLASS_ITERATORS
#undef CONTAINER
void f() {
using namespace std;
vector<int> std_vec;
vector<int>::iterator I = std_vec.begin();
// CHECK: auto I = std_vec.begin();
}
> As for the matchers, hasStdContainerName and isFromStdNamespace can be
combined but I think it is a better idea to have them separated. What do you
think?
I think so too. Something similar will be needed for some future matchers
dealing with STL types, so I IMHO if `isFromStdNamespace()` doesn't fit we
should change it instead of re-implementing the logic.
Apart from the test case it looks good to me.
================
Comment at: cpp11-migrate/UseAuto/UseAutoMatchers.cpp:237
@@ -210,3 +236,3 @@
),
hasPrefix(
specifiesNamespace(hasName("std"))
----------------
I'm not sure if this part is still needed but I don't think it hurts either.
http://llvm-reviews.chandlerc.com/D1116
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
