[clang-format] Add BreakAdjacentStringLiterals option (#73432)

Closes #70451.
llvm · Nov 27, 2023 · 39faf13 · 39faf13
1 parent 9e86919
commit 39faf13
Show file tree

Hide file tree

Showing 7 changed files with 52 additions and 5 deletions.
diff --git a/clang/docs/ClangFormatStyleOptions.rst b/clang/docs/ClangFormatStyleOptions.rst
@@ -2046,6 +2046,21 @@ the configuration (without a prefix: ``Auto``).
       };
     }
 
+.. _BreakAdjacentStringLiterals:
+
+**BreakAdjacentStringLiterals** (``Boolean``) :versionbadge:`clang-format 18` :ref:`¶ <BreakAdjacentStringLiterals>`
+  Break between adjacent string literals.
+
+  .. 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";
+
 .. _BreakAfterAttributes:
 
 **BreakAfterAttributes** (``AttributeBreakingStyle``) :versionbadge:`clang-format 16` :ref:`¶ <BreakAfterAttributes>`

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
@@ -956,6 +956,7 @@ clang-format
 - Add ``AllowBreakBeforeNoexceptSpecifier`` option.
 - Add ``AllowShortCompoundRequirementOnASingleLine`` option.
 - Change ``BreakAfterAttributes`` from ``Never`` to ``Leave`` in LLVM style.
+- Add ``BreakAdjacentStringLiterals`` option.
 
 libclang
 --------

diff --git a/clang/include/clang/Format/Format.h b/clang/include/clang/Format/Format.h
@@ -1424,6 +1424,19 @@ struct FormatStyle {
   /// \version 3.8
   BraceWrappingFlags BraceWrapping;
 
+  /// Break between adjacent string literals.
+  /// \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 BreakAdjacentStringLiterals;
+
   /// Different ways to break after attributes.
   enum AttributeBreakingStyle : int8_t {
     /// Always break after attributes.
@@ -4745,6 +4758,7 @@ struct FormatStyle {
            BinPackParameters == R.BinPackParameters &&
            BitFieldColonSpacing == R.BitFieldColonSpacing &&
            BracedInitializerIndentWidth == R.BracedInitializerIndentWidth &&
+           BreakAdjacentStringLiterals == R.BreakAdjacentStringLiterals &&
            BreakAfterAttributes == R.BreakAfterAttributes &&
            BreakAfterJavaFieldAnnotations == R.BreakAfterJavaFieldAnnotations &&
            BreakArrays == R.BreakArrays &&

diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp
@@ -965,6 +965,8 @@ template <> struct MappingTraits<FormatStyle> {
     IO.mapOptional("BracedInitializerIndentWidth",
                    Style.BracedInitializerIndentWidth);
     IO.mapOptional("BraceWrapping", Style.BraceWrapping);
+    IO.mapOptional("BreakAdjacentStringLiterals",
+                   Style.BreakAdjacentStringLiterals);
     IO.mapOptional("BreakAfterAttributes", Style.BreakAfterAttributes);
     IO.mapOptional("BreakAfterJavaFieldAnnotations",
                    Style.BreakAfterJavaFieldAnnotations);
@@ -1476,6 +1478,7 @@ FormatStyle getLLVMStyle(FormatStyle::LanguageKind Language) {
                              /*SplitEmptyFunction=*/true,
                              /*SplitEmptyRecord=*/true,
                              /*SplitEmptyNamespace=*/true};
+  LLVMStyle.BreakAdjacentStringLiterals = true;
   LLVMStyle.BreakAfterAttributes = FormatStyle::ABS_Leave;
   LLVMStyle.BreakAfterJavaFieldAnnotations = false;
   LLVMStyle.BreakArrays = true;

diff --git a/clang/lib/Format/TokenAnnotator.cpp b/clang/lib/Format/TokenAnnotator.cpp
@@ -5077,11 +5077,10 @@ 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 ||
-             Style.Language == FormatStyle::LK_Proto ||
-             Style.Language == FormatStyle::LK_TableGen ||
-             Style.Language == FormatStyle::LK_TextProto) {
+  } else if (Style.BreakAdjacentStringLiterals &&
+             (Style.isCpp() || Style.isProto() ||
+              Style.Language == FormatStyle::LK_TableGen ||
+              Style.Language == FormatStyle::LK_TextProto)) {
     if (Left.isStringLiteral() && Right.isStringLiteral())
       return true;
   }

diff --git a/clang/unittests/Format/ConfigParseTest.cpp b/clang/unittests/Format/ConfigParseTest.cpp
@@ -159,6 +159,7 @@ TEST(ConfigParseTest, ParsesConfigurationBools) {
   CHECK_PARSE_BOOL(AllowShortLoopsOnASingleLine);
   CHECK_PARSE_BOOL(BinPackArguments);
   CHECK_PARSE_BOOL(BinPackParameters);
+  CHECK_PARSE_BOOL(BreakAdjacentStringLiterals);
   CHECK_PARSE_BOOL(BreakAfterJavaFieldAnnotations);
   CHECK_PARSE_BOOL(BreakBeforeTernaryOperators);
   CHECK_PARSE_BOOL(BreakStringLiterals);

diff --git a/clang/unittests/Format/FormatTest.cpp b/clang/unittests/Format/FormatTest.cpp
@@ -26644,6 +26644,20 @@ TEST_F(FormatTest, StreamOutputOperator) {
   verifyFormat("std::cout << \"foo\" << \"bar\" << baz;");
 }
 
+TEST_F(FormatTest, BreakAdjacentStringLiterals) {
+  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.BreakAdjacentStringLiterals = false;
+  verifyFormat(Code, Style);
+}
 } // namespace
 } // namespace test
 } // namespace format