[clang] [clang-repl] Lay the basic infrastructure for pretty printing of types (PR #148701)

[Vassil Vassilev via cfe-commits]

================
@@ -0,0 +1,59 @@
+// RUN: clang-repl "int i = 10;" 'extern "C" int printf(const char*,...);' \
+// RUN:            'auto r1 = printf("i = %d\n", i);' | FileCheck --check-prefix=CHECK-DRIVER %s
+// UNSUPPORTED: system-aix
+// CHECK-DRIVER: i = 10
+// RUN: cat %s | clang-repl -Xcc -std=c++11 -Xcc -fno-delayed-template-parsing | FileCheck %s
+extern "C" int printf(const char*,...);
+
+"ab"
+// CHECK: (const char[3]) "ab"
+
+char ch[2] = {'1','a'}; ch
+// CHECK-NEXT: (char[2]) { '1', 'a' }
+
+char chnull[3] = {'1','a', '\0'}; chnull
+// CHECK-NEXT: (char[3]) "1a"
+
+char ch_arr[2][3][1] = {{{'a'}, {'b'}, {'c'}}, {{'d'}, {'e'}, {'f'}}}; ch_arr
+// CHECK: (char[2][3][1]) { { { 'a' }, { 'b' }, { 'c' } }, { { 'd' }, { 'e' }, { 'f' } } }
+struct S3 { int* p; S3() { p = new int(42); } ~S3() { delete p; } };
+S3{}
+// CHECK-NEXT: (S3) @0x{{[0-9a-f]+}}
+S3 s3;
+s3
+// CHECK-NEXT: (S3 &) @0x{{[0-9a-f]+}}
+
+struct S4 { ~S4() { printf("~S4()\n"); }};
+S4{}
+// CHECK-NEXT: (S4) @0x{{[0-9a-f]+}}
+
+enum Enum{ e1 = -12, e2, e3=33, e4, e5 = 33};
+e2
+// CHECK-NEXT: (Enum) (e2) : int -11
+::e1
+// CHECK-NEXT: (Enum) (e1) : int -12
+
+enum class Color { R = 0, G, B };
+Color::R
+// CHECK-NEXT: (Color) (Color::R) : int 0
+
+
+// Lambdas.
+
+auto Lambda1 = []{};
+Lambda1
+// CHECK-NEXT: ((lambda) &) @0x{{[0-9a-f]+}}
+[]{}
+// CHECK-NEXT: ((lambda at input_line_{{[0-9]+}}:1:1)) @0x{{[0-9a-f]+}}
+
+template<int n> struct F{ enum {RET=F<n-1>::RET*n} ; };
+template<> struct F<0> { enum {RET = 1}; };
+F<7>::RET
+// CHECK-NEXT: (F<7>::(unnamed enum at input_line_{{[0-9]+}}:1:27)) (F<7>::RET) : unsigned int 5040
+
+struct S5 { int foo() { return 42; }};
+&S5::foo
+// CHECK-NEXT: (int (S5::*)()) Function @0x{{[0-9a-f]+}}
+
----------------
vgvassilev wrote:

In that case we will need to add more infrastructure such as something like `getEffectiveType` and rework the code in the places where we assume that we deal with builtin types and not references to builtin types.

That's on my todo list but I'd like to tackle this in a subsequent PR to avoid expanding the current one too much if that's okay with you.

https://github.com/llvm/llvm-project/pull/148701