2 files changed, 67 insertions, 3 deletions
diff --git a/clang/test/CodeGen/char-literal.c b/clang/test/CodeGen/char-literal.c
new file mode 100644
index 00000000000..aff76d280d3
--- /dev/null
+++ b/clang/test/CodeGen/char-literal.c
@@ -0,0 +1,35 @@
+// RUN: %clang_cc1 -x c++ -triple i386-unknown-unkown -emit-llvm %s -o - | FileCheck %s
+// Runs in c++ mode so that wchar_t is available.
+
+int main() {
+  // CHECK: store i8 97
+  char a = 'a';
+
+  // Should pick second character.
+  // CHECK: store i8 98
+  char b = 'ab';
+
+  // CHECK: store i32 97
+  wchar_t wa = L'a';
+
+  // Should pick second character.
+  // CHECK: store i32 98
+  wchar_t wb = L'ab';
+
+  // Should pick last character and store its lowest byte.
+  // This does not match gcc, which takes the last character, converts it to
+  // utf8, and then picks the second-lowest byte of that (they probably store
+  // the utf8 in uint16_ts internally and take the lower byte of that).
+  // CHECK: store i8 48
+  char c = '\u1120\u0220\U00102030';
+
+  // CHECK: store i32 61451
+  wchar_t wc = L'\uF00B';
+
+  // CHECK: store i32 1110027
+  wchar_t wd = L'\U0010F00B';
+
+  // Should pick second character.
+  // CHECK: store i32 1110027
+  wchar_t we = L'\u1234\U0010F00B';
+}
diff --git a/clang/test/CodeGen/string-literal-short-wstring.c b/clang/test/CodeGen/string-literal-short-wstring.c
index de84953dd3f..be1f1dd66c7 100644
--- a/clang/test/CodeGen/string-literal-short-wstring.c
+++ b/clang/test/CodeGen/string-literal-short-wstring.c
@@ -1,4 +1,5 @@
-// RUN: %clang_cc1 -emit-llvm -fshort-wchar %s -o - | FileCheck %s
+// RUN: %clang_cc1 -x c++ -emit-llvm -fshort-wchar %s -o - | FileCheck %s
+// Runs in c++ mode so that wchar_t is available.
 
 int main() {
   // This should convert to utf8.
@@ -6,9 +7,37 @@ int main() {
   char b[10] = "\u1120\u0220\U00102030";
 
   // CHECK: private constant [6 x i8] c"A\00B\00\00\00"
-  void *foo = L"AB";
+  const wchar_t *foo = L"AB";
 
   // This should convert to utf16.
   // CHECK: private constant [10 x i8] c" \11 \02\C8\DB0\DC\00\00"
-  void *bar = L"\u1120\u0220\U00102030";
+  const wchar_t *bar = L"\u1120\u0220\U00102030";
+
+
+
+  // Should pick second character.
+  // CHECK: store i8 98
+  char c = 'ab';
+
+  // CHECK: store i16 97
+  wchar_t wa = L'a';
+
+  // Should pick second character.
+  // CHECK: store i16 98
+  wchar_t wb = L'ab';
+
+  // -4085 == 0xf00b
+  // CHECK: store i16 -4085
+  wchar_t wc = L'\uF00B';
+
+  // Should take lower word of the 4byte UNC sequence. This does not match
+  // gcc. I don't understand what gcc does (it looks like it converts to utf16,
+  // then takes the second (!) utf16 word, swaps the lower two nibbles, and
+  // stores that?).
+  // CHECK: store i16 -4085
+  wchar_t wd = L'\U0010F00B';  // has utf16 encoding dbc8 dcb0
+
+  // Should pick second character. (gcc: -9205)
+  // CHECK: store i16 -4085
+  wchar_t we = L'\u1234\U0010F00B';
 }