[Tooling][Format] Treat compound extensions (foo.bar.cc) as matching foo.h

[sammccall]

Summary

Motivating use case is ".cu.cc" extensions used in some bazel projects.

Alternative is to work around this with IncludeIsMainRegex in styles.
I proposed this approach because it seems like a better default.

[MyDeveloperDay]

LGTM

clang/lib/Tooling/Inclusions/HeaderIncludes.cpp

 //===--- HeaderIncludes.cpp - Insert/Delete #includes --*- C++ -*----------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 #include "clang/Tooling/Inclusions/HeaderIncludes.h"
 #include "clang/Basic/FileManager.h"
 #include "clang/Basic/SourceManager.h"
 #include "clang/Lex/Lexer.h"
 #include "llvm/ADT/Optional.h"
 #include "llvm/Support/FormatVariadic.h"
+#include "llvm/Support/Path.h"
 
 namespace clang {
 namespace tooling {
 namespace {
 
 LangOptions createLangOpts() {
   LangOptions LangOpts;
   LangOpts.CPlusPlus = 1;

 
 inline StringRef trimInclude(StringRef IncludeName) {
   return IncludeName.trim("\"<>");
 }
 
 const char IncludeRegexPattern[] =
     R"(^[\t\ ]*#[\t\ ]*(import|include)[^"<]*(["<][^">]*[">]))";
 
+// The filename of Path excluding extension.
+// Used to match implementation with headers, this differs from sys::path::stem:
+//  - in names with multiple dots (foo.cu.cc) it terminates at the *first*
+//  - an empty stem is never returned: /foo/.bar.x => .bar
+//  - we don't bother to handle . and .. specially
+StringRef matchingStem(llvm::StringRef Path) {
+  StringRef Name = llvm::sys::path::filename(Path);
+  return Name.substr(0, Name.find('.', 1));
+}
+
 } // anonymous namespace
 
 IncludeCategoryManager::IncludeCategoryManager(const IncludeStyle &Style,
                                                StringRef FileName)
     : Style(Style), FileName(FileName) {
-  FileStem = llvm::sys::path::stem(FileName);
+  FileStem = matchingStem(FileName);
   for (const auto &Category : Style.IncludeCategories)
     CategoryRegexs.emplace_back(Category.Regex, llvm::Regex::IgnoreCase);
   IsMainFile = FileName.endswith(".c") || FileName.endswith(".cc") ||
                FileName.endswith(".cpp") || FileName.endswith(".c++") ||
                FileName.endswith(".cxx") || FileName.endswith(".m") ||
                FileName.endswith(".mm");
   if (!Style.IncludeIsMainSourceRegex.empty()) {
     llvm::Regex MainFileRegex(Style.IncludeIsMainSourceRegex);
for (unsigned i = 0, e = CategoryRegexs.size(); i != e; ++i)
     }
   if (CheckMainHeader && IsMainFile && Ret > 0 && isMainHeader(IncludeName))
     Ret = 0;
   return Ret;
 }
 bool IncludeCategoryManager::isMainHeader(StringRef IncludeName) const {
   if (!IncludeName.startswith("\""))
     return false;
-  StringRef HeaderStem =
-      llvm::sys::path::stem(IncludeName.drop_front(1).drop_back(1));
+  StringRef HeaderStem = matchingStem(IncludeName.drop_front(1).drop_back(1));
   if (FileStem.startswith(HeaderStem) ||
       FileStem.startswith_lower(HeaderStem)) {
     llvm::Regex MainIncludeRegex(HeaderStem.str() + Style.IncludeIsMainRegex,
                                  llvm::Regex::IgnoreCase);
     if (MainIncludeRegex.match(FileStem))
       return true;
   }
   return false;

clang/unittests/Tooling/HeaderIncludesTest.cpp

std::string remove(llvm::StringRef Code, llvm::StringRef Header) {
     HeaderIncludes Includes(FileName, Code, Style);
     assert(Header.startswith("\"") || Header.startswith("<"));
     auto Replaces = Includes.remove(Header.trim("\"<>"), Header.startswith("<"));
     auto Result = applyAllReplacements(Code, Replaces);
     EXPECT_TRUE(static_cast<bool>(Result));
     return *Result;
   }
 
-  const std::string FileName = "fix.cpp";
+  std::string FileName = "fix.cpp";
   IncludeStyle Style = format::getLLVMStyle().IncludeStyle;
 };
 
 TEST_F(HeaderIncludesTest, NoExistingIncludeWithoutDefine) {
   std::string Code = "int main() {}";
   std::string Expected = "#include \"a.h\"\n"
                          "int main() {}";
   EXPECT_EQ(Expected, insert(Code, "\"a.h\""));
std::string Code = "#include \"fix.h\"\n"
                      "int main() {}";
   std::string Expected = "#include \"fix.h\"\n"
                          "#include <a>\n"
                          "\n"
                          "int main() {}";
   Style = format::getGoogleStyle(format::FormatStyle::LanguageKind::LK_Cpp)
               .IncludeStyle;
   EXPECT_EQ(Expected, insert(Code, "<a>"));
+
+  FileName = "fix.cu.cpp";
+  EXPECT_EQ(Expected, insert(Code, "<a>"));
+
+  FileName = "fix_test.cu.cpp";
+  EXPECT_EQ(Expected, insert(Code, "<a>"));
+
+  FileName = "bar.cpp";
+  EXPECT_NE(Expected, insert(Code, "<a>")) << "Not main header";
 }
 
 TEST_F(HeaderIncludesTest, InsertBeforeSystemHeaderLLVM) {
   std::string Code = "#include <memory>\n"
                      "\n"
                      "int main() {}";
   std::string Expected = "#include \"z.h\"\n"
                          "#include <memory>\n"