21041c9292
Most of the cases were where a C++ file was being compiled with the C substitution. There were a few cases of the opposite though. LLDB seems to be the only real culprit in the LLVM codebase for these mismatches. Rest of the LLVM presumably sticks at least language-specific options in the common substitutions making the mistakes immediately apparent. I found these by using Clang frontend configuration files containing language-specific options for both C and C++ (e.g. `-std=c2y` and `-std=c++26`).
64 lines
2.9 KiB
C++
64 lines
2.9 KiB
C++
// REQUIRES: lld
|
|
|
|
// Itanium ABI:
|
|
// RUN: %clangxx --target=x86_64-pc-linux -gdwarf -c -o %t_linux.o %s
|
|
// RUN: %lldb -f %t_linux.o -b -o "target variable s1 s2 m1 m2 v1 v2 v3 v4" | FileCheck --check-prefix=CHECK-GNU %s
|
|
//
|
|
// CHECK-GNU: (void (Single1::*)()) s1 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Single2::*)()) s2 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Multiple1::*)()) m1 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Multiple2::*)()) m2 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Virtual1::*)()) v1 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Virtual2::*)()) v2 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Virtual3::*)()) v3 = 0x00000000000000000000000000000000
|
|
// CHECK-GNU: (void (Virtual4::*)()) v4 = 0x00000000000000000000000000000000
|
|
|
|
// Microsoft ABI:
|
|
// RUN: %clang_cl --target=x86_64-windows-msvc -c -gdwarf -o %t_win.obj /GS- -- %s
|
|
// RUN: lld-link /out:%t_win.exe %t_win.obj /entry:main /debug /nodefaultlib
|
|
// RUN: %lldb -f %t_win.exe -b -o "target variable s1 s2 m1 m2 v1 v2 v3 v4" | FileCheck --check-prefix=CHECK-MSVC %s
|
|
//
|
|
// CHECK-MSVC: (void (Single1::*)()) s1 = 0x0000000000000000
|
|
// CHECK-MSVC: (void (Single2::*)()) s2 = 0x0000000000000000
|
|
// CHECK-MSVC: (void (Multiple1::*)()) m1 = 0x00000000000000000000000000000000
|
|
// CHECK-MSVC: (void (Multiple2::*)()) m2 = 0x00000000000000000000000000000000
|
|
// CHECK-MSVC: (void (Virtual1::*)()) v1 = 0xffffffff000000000000000000000000
|
|
// CHECK-MSVC: (void (Virtual2::*)()) v2 = 0xffffffff000000000000000000000000
|
|
// CHECK-MSVC: (void (Virtual3::*)()) v3 = 0xffffffff000000000000000000000000
|
|
// CHECK-MSVC: (void (Virtual4::*)()) v4 = 0xffffffff000000000000000000000000
|
|
|
|
// clang-format off
|
|
struct Single1 { void s1() {} };
|
|
struct Single2 : Single1 { void s2() {} };
|
|
|
|
struct Helper {};
|
|
struct Multiple1 : Single1, Helper { void m1() {} };
|
|
struct Multiple2 : Multiple1 { void m2() {} };
|
|
|
|
struct Virtual1 : virtual Single1 { void v1() {} };
|
|
struct Virtual2 : Virtual1 { void v2() {} };
|
|
struct Virtual3 : virtual Multiple1 { void v3() {} };
|
|
struct Virtual4 : Virtual1, Helper { void v4() {} };
|
|
|
|
void (Single1::*s1)() = nullptr;
|
|
void (Single2::*s2)() = nullptr;
|
|
void (Multiple1::*m1)() = nullptr;
|
|
void (Multiple2::*m2)() = nullptr;
|
|
void (Virtual1::*v1)() = nullptr;
|
|
void (Virtual2::*v2)() = nullptr;
|
|
void (Virtual3::*v3)() = nullptr;
|
|
void (Virtual4::*v4)() = nullptr;
|
|
|
|
int main(int argc, char *argv[]) {
|
|
// We need to force emission of type info to DWARF. That's reasonable, because
|
|
// Clang doesn't know that we need it to infer member-pointer sizes. We could
|
|
// probably teach Clang to do so, but in most real-world scenarios this might
|
|
// be a non-issue.
|
|
Virtual1 vi1;
|
|
Virtual2 vi2;
|
|
Virtual3 vi3;
|
|
Virtual4 vi4;
|
|
int sum = sizeof(Single2) + sizeof(Multiple2);
|
|
return argc < sum ? 0 : 1;
|
|
}
|