[clang] [clang-format] Add BreakConcatenatedStrings option (PR #73432)

[via cfe-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-clang-format

Author: Owen Pan (owenca)

<details>
<summary>Changes</summary>

Closes #<!-- -->70451.

---
Full diff: https://github.com/llvm/llvm-project/pull/73432.diff


7 Files Affected:

- (modified) clang/docs/ClangFormatStyleOptions.rst (+15) 
- (modified) clang/docs/ReleaseNotes.rst (+1) 
- (modified) clang/include/clang/Format/Format.h (+14) 
- (modified) clang/lib/Format/Format.cpp (+2) 
- (modified) clang/lib/Format/TokenAnnotator.cpp (+1-2) 
- (modified) clang/unittests/Format/ConfigParseTest.cpp (+1) 
- (modified) clang/unittests/Format/FormatTest.cpp (+14) 


``````````diff
diff --git a/clang/docs/ClangFormatStyleOptions.rst b/clang/docs/ClangFormatStyleOptions.rst
index ff424828ff63cdb..4ff9293b3de7a03 100644
--- a/clang/docs/ClangFormatStyleOptions.rst
+++ b/clang/docs/ClangFormatStyleOptions.rst
@@ -2761,6 +2761,21 @@ the configuration (without a prefix: ``Auto``).
          firstValue :
          SecondValueVeryVeryVeryVeryLong;
 
+.. _BreakConcatenatedStrings:
+
+**BreakConcatenatedStrings** (``Boolean``) :versionbadge:`clang-format 18` :ref:`¶ <BreakConcatenatedStrings>`
+  Break between concatenated string literals in C, C++, and Objective-C.
+
+  .. code-block:: c++
+
+     true:
+     return "Code"
+            "\0\52\26\55\55\0"
+            "x013"
+            "\02\xBA";
+     false:
+     return "Code" "\0\52\26\55\55\0" "x013" "\02\xBA";
+
 .. _BreakConstructorInitializers:
 
 **BreakConstructorInitializers** (``BreakConstructorInitializersStyle``) :versionbadge:`clang-format 5` :ref:`¶ <BreakConstructorInitializers>`
diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 52c9d6eb69617b0..3aa408496101ad3 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -884,6 +884,7 @@ clang-format
 - Add ``AllowBreakBeforeNoexceptSpecifier`` option.
 - Add ``AllowShortCompoundRequirementOnASingleLine`` option.
 - Change ``BreakAfterAttributes`` from ``Never`` to ``Leave`` in LLVM style.
+- Add ``BreakConcatenatedStrings`` option.
 
 libclang
 --------
diff --git a/clang/include/clang/Format/Format.h b/clang/include/clang/Format/Format.h
index bc412611ef62493..b5f7c16774cd96e 100644
--- a/clang/include/clang/Format/Format.h
+++ b/clang/include/clang/Format/Format.h
@@ -2088,6 +2088,19 @@ struct FormatStyle {
   /// \version 3.7
   bool BreakBeforeTernaryOperators;
 
+  /// Break between concatenated string literals in C, C++, and Objective-C.
+  /// \code
+  ///    true:
+  ///    return "Code"
+  ///           "\0\52\26\55\55\0"
+  ///           "x013"
+  ///           "\02\xBA";
+  ///    false:
+  ///    return "Code" "\0\52\26\55\55\0" "x013" "\02\xBA";
+  /// \endcode
+  /// \version 18
+  bool BreakConcatenatedStrings;
+
   /// Different ways to break initializers.
   enum BreakConstructorInitializersStyle : int8_t {
     /// Break constructor initializers before the colon and after the commas.
@@ -4753,6 +4766,7 @@ struct FormatStyle {
            BreakBeforeConceptDeclarations == R.BreakBeforeConceptDeclarations &&
            BreakBeforeInlineASMColon == R.BreakBeforeInlineASMColon &&
            BreakBeforeTernaryOperators == R.BreakBeforeTernaryOperators &&
+           BreakConcatenatedStrings == R.BreakConcatenatedStrings &&
            BreakConstructorInitializers == R.BreakConstructorInitializers &&
            BreakInheritanceList == R.BreakInheritanceList &&
            BreakStringLiterals == R.BreakStringLiterals &&
diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp
index e1abcac5604a410..025bca2ea9f7365 100644
--- a/clang/lib/Format/Format.cpp
+++ b/clang/lib/Format/Format.cpp
@@ -978,6 +978,7 @@ template <> struct MappingTraits<FormatStyle> {
                    Style.BreakBeforeInlineASMColon);
     IO.mapOptional("BreakBeforeTernaryOperators",
                    Style.BreakBeforeTernaryOperators);
+    IO.mapOptional("BreakConcatenatedStrings", Style.BreakConcatenatedStrings);
     IO.mapOptional("BreakConstructorInitializers",
                    Style.BreakConstructorInitializers);
     IO.mapOptional("BreakInheritanceList", Style.BreakInheritanceList);
@@ -1485,6 +1486,7 @@ FormatStyle getLLVMStyle(FormatStyle::LanguageKind Language) {
   LLVMStyle.BreakBeforeInlineASMColon = FormatStyle::BBIAS_OnlyMultiline;
   LLVMStyle.AllowBreakBeforeNoexceptSpecifier = FormatStyle::BBNSS_Never;
   LLVMStyle.BreakBeforeTernaryOperators = true;
+  LLVMStyle.BreakConcatenatedStrings = true;
   LLVMStyle.BreakConstructorInitializers = FormatStyle::BCIS_BeforeColon;
   LLVMStyle.BreakInheritanceList = FormatStyle::BILS_BeforeColon;
   LLVMStyle.BreakStringLiterals = true;
diff --git a/clang/lib/Format/TokenAnnotator.cpp b/clang/lib/Format/TokenAnnotator.cpp
index cfca7c00312aab3..17d3b48e0af8332 100644
--- a/clang/lib/Format/TokenAnnotator.cpp
+++ b/clang/lib/Format/TokenAnnotator.cpp
@@ -5078,8 +5078,7 @@ bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
     // it is hard to identify them in UnwrappedLineParser.
     if (!Keywords.isVerilogBegin(Right) && Keywords.isVerilogEndOfLabel(Left))
       return true;
-  } else if (Style.Language == FormatStyle::LK_Cpp ||
-             Style.Language == FormatStyle::LK_ObjC ||
+  } else if ((Style.isCpp() && Style.BreakConcatenatedStrings) ||
              Style.Language == FormatStyle::LK_Proto ||
              Style.Language == FormatStyle::LK_TableGen ||
              Style.Language == FormatStyle::LK_TextProto) {
diff --git a/clang/unittests/Format/ConfigParseTest.cpp b/clang/unittests/Format/ConfigParseTest.cpp
index 3a6667dbe0eb15d..910f5157a4f0209 100644
--- a/clang/unittests/Format/ConfigParseTest.cpp
+++ b/clang/unittests/Format/ConfigParseTest.cpp
@@ -161,6 +161,7 @@ TEST(ConfigParseTest, ParsesConfigurationBools) {
   CHECK_PARSE_BOOL(BinPackParameters);
   CHECK_PARSE_BOOL(BreakAfterJavaFieldAnnotations);
   CHECK_PARSE_BOOL(BreakBeforeTernaryOperators);
+  CHECK_PARSE_BOOL(BreakConcatenatedStrings);
   CHECK_PARSE_BOOL(BreakStringLiterals);
   CHECK_PARSE_BOOL(CompactNamespaces);
   CHECK_PARSE_BOOL(DerivePointerAlignment);
diff --git a/clang/unittests/Format/FormatTest.cpp b/clang/unittests/Format/FormatTest.cpp
index 67423f9b06fbc37..f83234e39ea036e 100644
--- a/clang/unittests/Format/FormatTest.cpp
+++ b/clang/unittests/Format/FormatTest.cpp
@@ -26642,6 +26642,20 @@ TEST_F(FormatTest, StreamOutputOperator) {
   verifyFormat("std::cout << \"foo\" << \"bar\" << baz;");
 }
 
+TEST_F(FormatTest, BreakConcatenatedStrings) {
+  constexpr StringRef Code{
+      "return \"Code\" \"\\0\\52\\26\\55\\55\\0\" \"x013\" \"\\02\\xBA\";"};
+
+  verifyFormat("return \"Code\"\n"
+               "       \"\\0\\52\\26\\55\\55\\0\"\n"
+               "       \"x013\"\n"
+               "       \"\\02\\xBA\";",
+               Code);
+
+  auto Style = getLLVMStyle();
+  Style.BreakConcatenatedStrings = false;
+  verifyFormat(Code, Style);
+}
 } // namespace
 } // namespace test
 } // namespace format

``````````

</details>


https://github.com/llvm/llvm-project/pull/73432