[clang-format] Don't sort qualifiers across preprocessor directives

[owenca]

Fixes #80579.

[llvmbot]

@llvm/pr-subscribers-clang-format

Author: Owen Pan (owenca)

Changes

Fixes #80579.

---

Full diff: https://github.com/llvm/llvm-project/pull/81958.diff

2 Files Affected:

• (modified) clang/lib/Format/QualifierAlignmentFixer.cpp (+2)
• (modified) clang/unittests/Format/QualifierFixerTest.cpp (+13)

diff --git a/clang/lib/Format/QualifierAlignmentFixer.cpp b/clang/lib/Format/QualifierAlignmentFixer.cpp
index 84941746f0df71..0c63d330b5aed4 100644
--- a/clang/lib/Format/QualifierAlignmentFixer.cpp
+++ b/clang/lib/Format/QualifierAlignmentFixer.cpp
@@ -561,6 +561,8 @@ void LeftRightQualifierAlignmentFixer::fixQualifierAlignment(
 
     for (const auto *Tok = First; Tok && Tok != Last && Tok->Next;
          Tok = Tok->Next) {
+      if (Tok->MustBreakBefore)
+        break;
       if (Tok->is(tok::comment))
         continue;
       if (RightAlign) {
diff --git a/clang/unittests/Format/QualifierFixerTest.cpp b/clang/unittests/Format/QualifierFixerTest.cpp
index 0aa755acfc8213..43476aea66337b 100644
--- a/clang/unittests/Format/QualifierFixerTest.cpp
+++ b/clang/unittests/Format/QualifierFixerTest.cpp
@@ -1177,6 +1177,19 @@ TEST_F(QualifierFixerTest, DontPushQualifierThroughNonSpecifiedTypes) {
   verifyFormat("const inline static Foo;", Style);
 }
 
+TEST_F(QualifierFixerTest, QualifiersBrokenUpByPPDirectives) {
+  auto Style = getLLVMStyle();
+  Style.QualifierAlignment = FormatStyle::QAS_Custom;
+  Style.QualifierOrder = {"constexpr", "inline", "type"};
+
+  verifyFormat("inline\n"
+               "#if FOO\n"
+               "    constexpr\n"
+               "#endif\n"
+               "    int i = 0;",
+               Style);
+}
+
 TEST_F(QualifierFixerTest, UnsignedQualifier) {
 
   FormatStyle Style = getLLVMStyle();

[mydeveloperday]

LGTM, as long as someone doesn't say it should be

#if FOO
   constexpr
#endif
   inline
   int i = 0;

but I think this change is fine..I would rather we didn't try and support that.. thank you.

[owenca]

LGTM, as long as someone doesn't say it should be

#if FOO
   constexpr
#endif
   inline
   int i = 0;

but I think this change is fine..I would rather we didn't try and support that.. thank you.

Yep! Otherwise, you would have to handle the following, given that the order is [const, inline, constexpr, type]:

inline
#if FOO
    constexpr /* foo */
#else
    const /* bar */
#endif
    int i = 0;