simpler workflow for -O gen-C++ ; also some hooks for -O gen-standalone-C++

src/script_opt/CPP/Compile.h

@@ -134,8 +134,8 @@ namespace zeek::detail {
 class CPPCompile {
 public:
 	CPPCompile(std::vector<FuncInfo>& _funcs, ProfileFuncs& pfs,
-	           const char* gen_name, CPPHashManager& hm,
-	           bool update, bool standalone);
+	           const std::string& gen_name, const std::string& addl_name,
+	           CPPHashManager& _hm, bool _update, bool _standalone);
 	~CPPCompile();
 
 private:
@@ -390,9 +390,14 @@ private:
 	// function.
 	std::string GenArgs(const RecordTypePtr& params, const Expr* e);
 
-	// Functions that we've declared/compiled.
+	// Functions that we've declared/compiled.  Indexed by full C++ name.
 	std::unordered_set<std::string> compiled_funcs;
 
+	// "Simple" functions that we've compiled, i.e., those that have
+	// a single body and thus can be called dirctly.  Indexed by
+	// function name, and maps to the C++ name.
+	std::unordered_map<std::string, std::string> compiled_simple_funcs;
+
 	// Maps those to their associated files - used to make add-C++ body
 	// hashes distinct.
 	std::unordered_map<std::string, std::string> cf_locs;
@@ -858,6 +863,12 @@ private:
 	void AddInit(const IntrusivePtr<Obj>& o)	{ AddInit(o.get()); }
 	void AddInit(const Obj* o);
 
+	// This is akin to an initialization, but done separately
+	// (upon "activation") so it can include initializations that
+	// rely on parsing having finished (in particular, BiFs having
+	// been registered).  Only used when generating  standalone code.
+	void AddActivation(std::string a)	{ activations.emplace_back(a); }
+
 	// Records the fact that the initialization of object o1 depends
 	// on that of object o2.
 	void NoteInitDependency(const IntrusivePtr<Obj>& o1,
@@ -922,6 +933,10 @@ private:
 	// other initializations, and that themselves have no dependencies).
 	std::vector<std::string> pre_inits;
 
+	// A list of "activations" (essentially, post-initializations).
+	// See AddActivation() above.
+	std::vector<std::string> activations;
+
 	// Expressions for which we need to generate initialization-time
 	// code.  Currently, these are only expressions appearing in
 	// attributes.
@@ -1010,6 +1025,9 @@ private:
 	// File to which we're generating code.
 	FILE* write_file;
 
+	// Name of file holding potential "additional" code.
+	std::string addl_name;
+
 	// Indentation level.
 	int block_level = 0;
 

src/script_opt/CPP/DeclFunc.cc

@@ -24,6 +24,9 @@ void CPPCompile::DeclareFunc(const FuncInfo& func)
 
 	DeclareSubclass(f->GetType(), pf, fname, body, priority, nullptr,
 	                f->Flavor());
+
+	if ( f->GetBodies().size() == 1 )
+		compiled_simple_funcs[f->Name()] = fname;
 	}
 
 void CPPCompile::DeclareLambda(const LambdaExpr* l, const ProfileFunc* pf)

src/script_opt/CPP/Driver.cc

@@ -13,20 +13,24 @@ using namespace std;
 
 
 CPPCompile::CPPCompile(vector<FuncInfo>& _funcs, ProfileFuncs& _pfs,
-                       const char* gen_name, CPPHashManager& _hm,
-                       bool _update, bool _standalone)
-: funcs(_funcs), pfs(_pfs), hm(_hm), update(_update), standalone(_standalone)
+                       const string& gen_name, const string& _addl_name,
+                       CPPHashManager& _hm, bool _update, bool _standalone)
+: funcs(_funcs), pfs(_pfs), hm(_hm),
+  update(_update), standalone(_standalone)
 	{
-	auto mode = hm.IsAppend() ? "a" : "w";
+	addl_name = _addl_name;
+	bool is_addl = hm.IsAppend();
+	auto target_name = is_addl ? addl_name.c_str() : gen_name.c_str();
+	auto mode = is_addl ? "a" : "w";
 
-	write_file = fopen(gen_name, mode);
+	write_file = fopen(target_name, mode);
 	if ( ! write_file )
 		{
-		reporter->Error("can't open C++ target file %s", gen_name);
+		reporter->Error("can't open C++ target file %s", target_name);
 		exit(1);
 		}
 
-	if ( hm.IsAppend() )
+	if ( is_addl )
 		{
 		// We need a unique number to associate with the name
 		// space for the code we're adding.  A convenient way to
@@ -39,7 +43,7 @@ CPPCompile::CPPCompile(vector<FuncInfo>& _funcs, ProfileFuncs& _pfs,
 			{
 			char buf[256];
 			util::zeek_strerror_r(errno, buf, sizeof(buf));
-			reporter->Error("fstat failed on %s: %s", gen_name, buf);
+			reporter->Error("fstat failed on %s: %s", target_name, buf);
 			exit(1);
 			}
 
@@ -49,6 +53,20 @@ CPPCompile::CPPCompile(vector<FuncInfo>& _funcs, ProfileFuncs& _pfs,
 		addl_tag = st.st_size + 1;
 		}
 
+	else
+		{
+		// Create an empty "additional" file.
+		auto addl_f = fopen(addl_name.c_str(), "w");
+		if ( ! addl_f )
+			{
+			reporter->Error("can't open C++ additional file %s",
+			                addl_name.c_str());
+			exit(1);
+			}
+
+		fclose(addl_f);
+		}
+
 	Compile();
 	}
 
@@ -313,7 +331,7 @@ void CPPCompile::GenEpilog()
 	if ( addl_tag > 0 )
 		return;
 
-	Emit("#include \"CPP-gen-addl.h\"\n");
+	Emit("#include \"" + addl_name + "\"\n");
 	Emit("} // zeek::detail");
 	}
 

src/script_opt/ScriptOpt.cc

@@ -280,9 +280,6 @@ void analyze_scripts()
 		// Avoid profiling overhead.
 		return;
 
-	const auto hash_name = hash_dir + "CPP-hashes";
-	const auto gen_name = hash_dir + "CPP-gen-addl.h";
-
 	// Now that everything's parsed and BiF's have been initialized,
 	// profile the functions.
 	auto pfs = std::make_unique<ProfileFuncs>(funcs, is_CPP_compilable, false);
@@ -376,6 +373,8 @@ void analyze_scripts()
 
 	if ( generating_CPP )
 		{
+		const auto hash_name = hash_dir + "CPP-hashes";
+
 		auto hm = std::make_unique<CPPHashManager>(hash_name.c_str(),
 						analysis_options.add_CPP);
 
@@ -394,7 +393,10 @@ void analyze_scripts()
 			pfs = std::make_unique<ProfileFuncs>(funcs, is_CPP_compilable, false);
 			}
 
-		CPPCompile cpp(funcs, *pfs, gen_name.c_str(), *hm,
+		const auto gen_name = hash_dir + "CPP-gen.cc";
+		const auto addl_name = hash_dir + "CPP-gen-addl.h";
+
+		CPPCompile cpp(funcs, *pfs, gen_name, addl_name, *hm,
 			       analysis_options.gen_CPP ||
 			       analysis_options.update_CPP,
 			       analysis_options.gen_standalone_CPP);