[clang-format] Fix designated initializer detection

Currently, in the following snippet, the second designated initializer
is incorrectly detected as an OBJC method expr. Fix that and a test to
make sure we don't regress.

```
Foo foo[] = {[0] = 1, [1] = 2};
```

Author: DaanDeMeyer

clang/lib/Format/TokenAnnotator.cpp

@@ -708,6 +708,11 @@ class AnnotatingParser {
         IsCpp && !IsCpp11AttributeSpecifier && !IsCSharpAttributeSpecifier &&
         Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) &&
         CurrentToken->isNoneOf(tok::l_brace, tok::r_square) &&
+        // Do not consider '[' after a comma inside a braced initializer the
+        // start of an ObjC method expression. In braced initializer lists,
+        // commas are list separators and should not trigger ObjC parsing.
+        (!Parent || Parent->is(tok::comma) ||
+         Contexts.back().ContextKind != tok::l_brace) &&
         (!Parent ||
          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
                          tok::kw_return, tok::kw_throw) ||

clang/unittests/Format/TokenAnnotatorTest.cpp

@@ -3361,6 +3361,11 @@ TEST_F(TokenAnnotatorTest, UnderstandDesignatedInitializers) {
   ASSERT_EQ(Tokens.size(), 14u) << Tokens;
   EXPECT_TOKEN(Tokens[6], tok::l_square, TT_DesignatedInitializerLSquare);
   EXPECT_BRACE_KIND(Tokens[9], BK_BracedInit);
+
+  Tokens = annotate("Foo foo[] = {[0] = 1, [1] = 2};");
+  ASSERT_EQ(Tokens.size(), 20u) << Tokens;
+  EXPECT_TOKEN(Tokens[6], tok::l_square, TT_DesignatedInitializerLSquare);
+  EXPECT_TOKEN(Tokens[12], tok::l_square, TT_DesignatedInitializerLSquare);
 }
 
 TEST_F(TokenAnnotatorTest, UnderstandsJavaScript) {