Implement detection of resource and include dirs.

This is useful when deployed CppInterOp needs to create an interpreter. Often
the interpreter itself cannot use the clang toolchain logic to discover both
the resource-dir and the include paths. This is because CppInterOp is a library
and usually put in the `lib` folder whereas the toolchain logic expects things
to be relative to the `bin` folder.

In such setups, we can ask CppInterOp to detect the resource directory by asking
clang about it. In the same vain, of finding C and C++ headers we can ask the
system compiler about the include paths.

Author: vgvassilev

include/clang/Interpreter/CppInterOp.h

@@ -444,6 +444,21 @@ namespace Cpp {
   /// Returns the resource-dir path (for headers).
   const char* GetResourceDir();
 
+  /// Uses the underlying clang compiler to detect the resource directory.
+  /// In essence calling clang -print-resource-dir and checks if it ends with
+  /// a compatible to CppInterOp version.
+  ///\param[in] ClangBinaryName - the name (or the full path) of the compiler
+  ///                             to ask.
+  std::string DetectResourceDir(const char* ClangBinaryName = "clang");
+
+  /// Asks the system compiler for its default include paths.
+  ///\param[out] Paths - the list of include paths returned by eg.
+  ///                     `c++ -xc++ -E -v /dev/null 2>&1`
+  ///\param[in] CompilerName - the name (or the full path) of the compiler
+  ///                          binary file.
+  void DetectSystemCompilerIncludePaths(std::vector<std::string>& Paths,
+                                        const char* CompilerName = "c++");
+
   /// Secondary search path for headers, if not found using the
   /// GetResourceDir() function.
   void AddIncludePath(const char *dir);

lib/Interpreter/CppInterOp.cpp

@@ -38,6 +38,10 @@
 #include <io.h>
 #ifndef STDOUT_FILENO
 #define STDOUT_FILENO 1
+// For exec().
+#include <stdio.h>
+#define popen(x, y) (_popen(x, y))
+#define pclose (_pclose)
 #endif
 #else
 #include <dlfcn.h>
@@ -2534,6 +2538,64 @@ namespace Cpp {
     return getInterp().getCI()->getHeaderSearchOpts().ResourceDir.c_str();
   }
 
+  ///\returns 0 on success.
+  static bool exec(const char* cmd, std::vector<std::string>& outputs) {
+#define DEBUG_TYPE "exec"
+
+    std::array<char, 256> buffer;
+    std::unique_ptr<FILE, decltype(&pclose)> pipe(popen(cmd, "r"), pclose);
+    LLVM_DEBUG(dbgs() << "Executing command '" << cmd << "'\n");
+
+    if (!pipe) {
+      LLVM_DEBUG(dbgs() << "Execute failed!\n");
+      perror("exec: ");
+      return false;
+    }
+
+    LLVM_DEBUG(dbgs() << "Execute returned:\n");
+    while (fgets(buffer.data(), static_cast<int>(buffer.size()), pipe.get())) {
+      LLVM_DEBUG(dbgs() << buffer.data());
+      llvm::StringRef trimmed = buffer.data();
+      outputs.push_back(trimmed.trim().str());
+    }
+
+#undef DEBUG_TYPE
+
+    return true;
+  }
+
+  std::string DetectResourceDir(const char* ClangBinaryName /* = clang */) {
+    std::string cmd = std::string(ClangBinaryName) + " -print-resource-dir";
+    std::vector<std::string> outs;
+    exec(cmd.c_str(), outs);
+    if (outs.empty() || outs.size() > 1)
+      return "";
+
+    std::string detected_resource_dir = outs.back();
+
+    std::string version =
+#if CLANG_VERSION_MAJOR < 16
+        CLANG_VERSION_STRING;
+#else
+        CLANG_VERSION_MAJOR_STRING;
+#endif
+    // We need to check if the detected resource directory is compatible.
+    if (llvm::sys::path::filename(detected_resource_dir) != version)
+      return "";
+
+    return detected_resource_dir;
+  }
+
+  void DetectSystemCompilerIncludePaths(std::vector<std::string>& Paths,
+                                        const char* CompilerName /*= "c++"*/) {
+    std::string cmd = "LC_ALL=C ";
+    cmd += CompilerName;
+    cmd += " -xc++ -E -v /dev/null 2>&1 | sed -n -e '/^.include/,${' -e '/^ "
+           "\\/.*/p' -e '}'";
+    std::vector<std::string> outs;
+    exec(cmd.c_str(), Paths);
+  }
+
   void AddIncludePath(const char *dir) {
     getInterp().AddIncludePath(dir);
   }

unittests/CppInterOp/CMakeLists.txt

@@ -18,6 +18,9 @@ target_link_libraries(CppInterOpTests
   clangCppInterOp
   )
 
+set_source_files_properties(InterpreterTest.cpp PROPERTIES COMPILE_DEFINITIONS
+  "LLVM_BINARY_DIR=\"${LLVM_BINARY_DIR}\""
+  )
 export_executable_symbols(CppInterOpTests)
 
 unset(LLVM_LINK_COMPONENTS)

unittests/CppInterOp/InterpreterTest.cpp

@@ -1,5 +1,8 @@
 #include "clang/Interpreter/CppInterOp.h"
 
+#include "llvm/ADT/SmallString.h"
+#include "llvm/Support/Path.h"
+
 #include <gmock/gmock.h>
 #include "gtest/gtest.h"
 
@@ -14,6 +17,7 @@ TEST(InterpreterTest, DISABLED_DebugFlag) {
 #else
 TEST(InterpreterTest, DebugFlag) {
 #endif // NDEBUG
+  Cpp::CreateInterpreter();
   EXPECT_FALSE(Cpp::IsDebugOutputEnabled());
   std::string cerrs;
   testing::internal::CaptureStderr();
@@ -49,7 +53,7 @@ TEST(InterpreterTest, Evaluate) {
   EXPECT_FALSE(HadError) ;
 }
 
-TEST(InterpreterTest, Process) {  
+TEST(InterpreterTest, Process) {
   Cpp::CreateInterpreter();
   EXPECT_TRUE(Cpp::Process("") == 0);
   EXPECT_TRUE(Cpp::Process("int a = 12;") == 0);
@@ -80,3 +84,22 @@ TEST(InterpreterTest, CreateInterpreter) {
   EXPECT_TRUE(Cpp::GetNamed("cpp17"));
   EXPECT_FALSE(Cpp::GetNamed("cppUnknown"));
 }
+
+#ifdef LLVM_BINARY_DIR
+TEST(InterpreterTest, DetectResourceDir) {
+#else
+TEST(InterpreterTest, DISABLED_DetectResourceDir) {
+#endif // LLVM_BINARY_DIR
+  Cpp::CreateInterpreter();
+  EXPECT_STRNE(Cpp::DetectResourceDir().c_str(), Cpp::GetResourceDir());
+  llvm::SmallString<256> Clang(LLVM_BINARY_DIR);
+  llvm::sys::path::append(Clang, "bin", "clang");
+  std::string DetectedPath = Cpp::DetectResourceDir(Clang.str().str().c_str());
+  EXPECT_STREQ(DetectedPath.c_str(), Cpp::GetResourceDir());
+}
+
+TEST(InterpreterTest, DetectSystemCompilerIncludePaths) {
+  std::vector<std::string> includes;
+  Cpp::DetectSystemCompilerIncludePaths(includes);
+  EXPECT_FALSE(includes.empty());
+}