path: root/cpp.cpp

#include "cpp.h"

#include "bnf.h"
#include "cppbnf.h"
#include "grammer.h"
#include "minicc.h"
#include "debug.h"

#include <gtest/gtest.h>
#include <gmock/gmock.h>

#include <unordered_set>

using namespace Gram;

CPP::CPP(){}

CPP::~CPP(){}

// Phase 1: Map physical character set to basic source character set
void CPP::source_charset_map()
{
 // TODO
}

// Phase 2: Escape backslashed line endings
void CPP::backslash_escape()
{
 // TODO
}

namespace {

std::vector<Token> sourceToCharTokens(const std::string& code)
{
 std::vector<Token> result;

 Location location{1, 1};

 for (char c: code) {
  if (c == '\n') {
   location.column = 1;
   location.line++;
  } else if (std::isprint(c)) {
   location.column++;
  }

  result.emplace_back(Token{std::string(1, c), std::string(1, c), location});
 }
 return result;
}

}

// Phase 3: Parse preprocessing tokens
std::pair<index_t, std::vector<TreeNode>> CPP::preprocessing_tokenize(const std::string& s)
{
 m_code = s;

 m_charTokens = sourceToCharTokens(s);

 auto bnf{SubBNF(GetCppBNFLex(), "preprocessing-token")};

 // add to bnf to match whole file
 bnf["file"] = {
  {"preprocessing-token-list"},
  {"whitespace-list", "preprocessing-token-list"}
 };
 bnf["preprocessing-token-list"] = {
  {"preprocessing-token-padded"},
  {"preprocessing-token-list", "preprocessing-token-padded"}
 };
 bnf["preprocessing-token-padded"] = {
  {"preprocessing-token"},
  {"preprocessing-token", "whitespace-list"}
 };
 bnf["whitespace-list"] = {
  {"whitespace-char"},
  {"whitespace-list", "whitespace-char" }
 };
 bnf["whitespace-char"] = {
  {" "}, {"\t"}, {"\n"}, {"\r"}
 };
 Gram::Compiler compiler(bnf, "file");
 debug = true;
 std::pair<index_t, std::vector<TreeNode>> Tree = compiler.compile(m_charTokens);

 debug = true;
 compiler.DumpTree();

 return Tree;
}

// Phase 4: Preprocessing
void CPP::preprocess()
{
 // TODO
}

// Phase 5: Map chars and strings to execution charset
void CPP::execution_charset_map()
{
 // TODO
}

// Phase 6: Concatenate adjacent string literals
void CPP::concatenate_strings()
{
 // TODO
}

std::string CPP::valueOfNode(index_t node_index, const std::vector<TreeNode>& Tree)
{
 std::string result;

 std::vector<int32_t> todo(1, int32_t(node_index));

 while (!todo.empty()) {
  int32_t current_index = todo.back();
  todo.pop_back();

  // visit node if token
  if (ChildIdIsToken(current_index)) {
   result += m_code[TokenIdFromChildId(current_index)];
  } else {

   const TreeNode &node{Tree[current_index]};

   // iterate backwards in childs, to get depth-first search in tree, from the beginning
   std::for_each(node.child_ids.rbegin(), node.child_ids.rend(), [&](int32_t child){
    todo.push_back(child);
   });
  }
 }

 return result;
};

namespace {
 std::unordered_set<std::string> pp_types{
  "identifier",
  "pp-number",
  "character-literal",
  "user-defined-character-literal",
  "string-literal",
  "user-defined-string-literal",
  "preprocessing-op-or-punc"
 };
}

// Phase 7.a: Create tokens from preprocessing tokens
std::vector<Token> CPP::tokens_from_pptokens(std::pair<index_t, std::vector<TreeNode>> Tree)
{
 std::vector<Token> result;

 // "identifier" + value -> "identifier" + value, except identifiers from table 5.11, p.14 -> keyword as value, value
 // "pp-number" + value -> "literal" + value
 // "character-literal" -> "literal" + value
 // "user-defined-character-literal" -> "literal" + value
 // "string-literal" -> "literal" + value
 // "user-defined-string-literal" -> "literal" + value
 // "preprocessing-op-or-punc" -> value+value (operator,punctuator)
 
 // TODO: traverse Tree, creating Token list
 std::vector<index_t> todo(1, index_t(Tree.first));

 while (!todo.empty()) {
  index_t current_index = todo.back();
  todo.pop_back();

  TreeNode &node{Tree.second[current_index]};

  // visit node
  if (pp_types.find(node.type) != pp_types.end()) { // TODO
   std::cout << node.type << ": " << valueOfNode(current_index, Tree.second) << std::endl;
  } else { // only traverse further if not handled

   // iterate backwards in childs, to get depth-first search in tree, from the beginning
   std::for_each(node.child_ids.rbegin(), node.child_ids.rend(), [&](int32_t child){
    if (!ChildIdIsToken(child))
     todo.push_back(child);
   });
  }
 }
 
 return result;
}

// Phase 7.b: Grammar Analysis
std::pair<index_t, std::vector<Gram::TreeNode>> analysis(std::vector<Token>)
{
 return {0 , {}};
}

// Phase 7.c: Translate
void CPP::translate()
{
 // TODO
}

// Phase 8: Instantiate objects
void CPP::instantiate()
{
 // TODO
}

// Phase 9: Link libraries
void CPP::link()
{
 // TODO
}

// phases of translation, according to standard
void CPP::translate(const std::string& code)
{
#if 0 // fix signatures!
 source_charset_map();
 backslash_escape();
 preprocessing_tokenize(code);
 preprocess();
 execution_charset_map();
 concatenate_strings();
 tokens_from_pptokens();
  analysis();
  translate();
 instantiate();
 link();
#endif
}

class CppTest: public ::testing::Test
{
protected:
 CppTest() {
  //debug = true;
 }
 ~CppTest() {
 }
};

#if 0
TEST_F(CppTest, preprocessing_tokenize) {
 CPP cpp;
 auto ppTree = cpp.preprocessing_tokenize("int main() { return 1; }");

 cpp.tokens_from_pptokens(ppTree);
}
#endif

#if 0
TEST_F(CppTest, preprocessing_tokenize2) {
 CPP cpp;
 auto ppTree = cpp.preprocessing_tokenize("in ma");

 cpp.tokens_from_pptokens(ppTree);
}
#endif

#if 0
TEST(Cpp, translate) {
 CPP::translate();
}
#endif