[PATCH] D43906: [clang-format] Improve detection of Objective-C block types

[Daniel Jasper via Phabricator via cfe-commits]

djasper added a comment.

Would it be enough to only add the block type case? With the macro false positive, there won't be an open paren after the closing paren, right?



================
Comment at: lib/Format/TokenAnnotator.cpp:155
+           Next->startsSequence(tok::identifier, tok::l_square,
+                                tok::numeric_constant, tok::r_square,
+                                tok::r_paren, tok::l_paren))) {
----------------
This seems suspect. Does it have to be a numeric_constant?


================
Comment at: unittests/Format/FormatTest.cpp:12029
+  EXPECT_EQ(FormatStyle::LK_Cpp, guessLanguage("foo.h", "FOO(^);"));
+  EXPECT_EQ(FormatStyle::LK_ObjC,
+            guessLanguage("foo.h", "int(^)(char, float);"));
----------------
I am somewhat skeptical about all these tests now being solely about guessLanguage. It might be the best choice for some of them, but also, there might be other things we do to detect ObjC at some point and then all of these tests become meaningless.

Does your change create a different formatting here?


Repository:
  rC Clang

https://reviews.llvm.org/D43906