C++语法扩展--简单的解释器

<C++编程艺术>之对于C++语法的拓展, 主要包括cases ... to ..., repeat..until..., foreach(...in...), breakon(...)等.

当然这一切都是不严格的, 只是demo一下而已, 把扩展后的C++代码解释成真正的C++代码, 没有做太多错误检查.

后来想了几个可扩展的点, 不过基于这段代码来实现的话会麻烦些, 比如用if else 模拟一个支持string类的switch, 对于switch的case里面的代码块用{}来wrap一下等,感觉使用递归下降的方法会很好搞~

下面是书里的代码:

// A translator for experimental C++ extensions.  
#include <iostream>  
#include <fstream>  
#include <cctype>  
#include <cstring>  
#include <string>  
  
using namespace std;  
  
// Prototypes for the functions that handle  
// the extended keywords.  
void foreach();  
void cases();  
void repeat();  
void until();  
void typeof();  
void breakon();

void ignore();
void elif();
  
// Prototypes for tokenizing the input file.  
bool gettoken(string &tok);  
void skipspaces();  
  
// Indentation padding string.  
string indent = "";  
  
// The input and output file streams.  
ifstream fin;  
ofstream fout;  
  
// Exception class for syntax errors.  
class SyntaxExc {  
  string what;  
public:  
  SyntaxExc(char *e) { what = string(e); }  
  string geterror() { return what; }  
};  
  
int main(int argc, char *argv[]) {  
  string token;  
  bool switch_start = true;
  
  if(argc != 3) {  
    cout << "Usage: ep <input file> <output file>\n";  
    return 1;  
  }  
  
  fin.open(argv[1]);  
  
  if(!fin) {  
    cout << "Cannot open " << argv[1] << endl;  
    return 1;  
  }  
  
  fout.open(argv[2]);  
  
  if(!fout) {  
    cout << "Cannot open " << argv[2] << endl;  
    return 1;  
  }  
  
  // Write header.  
  fout << "// Translated from an .exp source file.\n";  
  
  try {  //为什么要优先于foreach来处理这几个情况呢, 因为foreach有可能在comment里面, 这样就会造成错误的解释.
    // Main translation loop.  
    while(gettoken(token)) {  
  
      // Skip over // comments.  
      if(token == "//") {  
        do {  
          fout << token;  
          gettoken(token);  
        } while(token.find('\n') == string::npos);  
        fout << token;  
      }  
  
      // Skip over /* comments.  
      else if(token == "/*") {  
        do {  
          fout << token;  
          gettoken(token);  
        } while(token != "*/");  
        fout << token;  
      }  
  
      // Skip over quoted string.  
      else if(token == "\"") {  
        do {  
          fout << token;  
          gettoken(token);  
        } while(token != "\"");  
        fout << token;  
      }
      else if(token == "\'")
      {
          do 
          {
              fout << token;
              gettoken(token);
          } while (token != "\'");
          fout << token;
      }
  
      else if(token == "foreach") foreach();  
  
      else if(token == "cases") cases();  
  
      else if(token == "repeat") repeat();  
  
      else if(token == "until") until();  
  
      else if(token == "typeof") typeof();  

      else if(token == "breakon") breakon();
      
      else if(token == "ignore") ignore();

      else if(token == "elif") elif();
  
      else fout << token;  
    }  
  } catch(SyntaxExc exc) {  
    cout << exc.geterror() << endl;  
    return 1;  
  }  
  
  return 0;  
}  


// Get the next token from the input stream.  
bool gettoken(string &tok) {  
  char ch;  
  char ch2;  
  static bool trackIndent = true;  
  
  tok = "";  
  
  ch = fin.get();  
  
  // Check for EOF and return false if EOF  
  // is found.  
  if(!fin) return false;  
  
  // Read whitespace.  
  if(isspace(ch)) {    
    while(isspace(ch)) {  
      tok += ch;  
  
      // Reset indent counter with each new line.  
      if(ch == '\n') {  
        indent = "";  
        trackIndent = true;  
      }  
      else if(trackIndent) indent += ch;  
  
      ch = fin.get();  
    }  
    fin.putback(ch); //为什么要回退呢? 因为读到一个非空白的, 后面还会有用~~
    return true;  
  }  
  
  // Stop tracking indentation after encountering  
  // first non-whitespace character on a line.  
  trackIndent = false;  
  
  // Read an identifier or keyword.  
  if(isalpha(ch) || ch=='_') {  
    while(isalpha(ch) || isdigit(ch) || ch=='_') {  
      tok += ch;  
      ch = fin.get();  
    }  
    fin.putback(ch);  
    return true;  
  }  
  
  // Read a number.  
  if(isdigit(ch)) {  
    while(isdigit(ch) || ch=='.' ||  
          tolower(ch) == 'e' ||  
          ch == '-' || ch =='+') {  
      tok += ch;  
      ch = fin.get();  
    }  
    fin.putback(ch);  
    return true;  
  }  
  
  // Check for \"  
  if(ch == '\\') {  
    ch2 = fin.get();
    tok += ch;
    tok += ch2;
    return true;
    
  }  
  if(ch == '=')
  {
      ch2 = fin.get();
      if(ch2 == '=')
      {
          tok += "==";
      }
      else
      {
          tok += "=";
          fin.putback(ch2);
      }
      return true;
  }
  
  
  // Check for begin comment symbols.  
  if(ch == '/') {  
    tok += ch;  
    ch = fin.get();  
    if(ch == '/' || ch == '*') {  
      tok += ch;  
    }  
    else fin.putback(ch);  
    return true;  
  }  
  
  // Check for end comment symbols.  
  if(ch == '*') {  
    tok += ch;  
    ch = fin.get();  
    if(ch == '/') {  
      tok += ch;  
    }  
    else fin.putback(ch);  
    return true;  
  }  
  
  tok += ch;  
  
  return true;  
}   
  
// Translate a foreach loop.  
void foreach() {  
  static bool flag = false;
  string token;  
  string varname;  
  string arrayname;  
  
  char forvarname[20] = "_foreach_idx_";  
  //static char counter[2] = "a";  
  static int counter = 0;
  char buf[50];
  sprintf(buf, "%d", counter);
  
  // Create loop control variable for generated  
  // for loop.  
  
  strcat(forvarname, buf);

  counter++;

  //counter[0]++;  
  
  // Only 26 foreach loops in a file because  
  // generated loop control variables limited to  
  // _ia to _iz. This can be changed if desired.  
  //if(counter[0] > 'z')   
    //throw SyntaxExc("Too many foreach loops.");  
  
  fout << "int " << forvarname  
       << " = 0;\n";  
  
  // Write beginning of generated for loop.  
  fout << indent << "for(";  
  
  skipspaces();  
  
  // Read the (  
  gettoken(token);  
  if(token[0] != '(')   
    throw SyntaxExc("( expected in foreach.");  
  
  skipspaces();  
  
  // Get the type of the foreach variable.   
  gettoken(token);  
  fout << token << " ";  
  
  skipspaces();  
  
  // Read and save the foreach variable's name.  
  gettoken(token);  
  varname = token;  
  
  skipspaces();  
  
  // Read the "in"  
  gettoken(token);  
  if(token != "in")   
    throw SyntaxExc("in expected in foreach.");  
  
  skipspaces();  
  
  // Read the array name.  
  gettoken(token);  
  arrayname = token;  
  
  fout << varname << " = " << arrayname << "[0];\n";  
  
  // Construct target value.  
  fout << indent + "    " << forvarname << " < "  
       << "((sizeof " << token << ")/"   
       << "(sizeof " << token << "[0]));\n";  
  
  fout << indent + "    " << forvarname << "++, "  
       << varname << " = " << arrayname << "["  
       << forvarname << "])";  
  
  skipspaces();  
  
  // Read the )  
  gettoken(token);  
  if(token[0] != ')')  
    throw SyntaxExc(") expected in foreach.");  
}  
  
// Translate a cases statement.  
void cases() {  
  string token;  
  int start, end;  
  
  skipspaces();  
  
  // Get starting value.  
  gettoken(token);  
  
  if(isdigit(token[0])) {  
    // is an int constant  
    start = atoi(token.c_str());  
  }  
  else if(token[0] == '\'') {  
    // is char constant  
    gettoken(token);  
    if(token.size() != 1) throw SyntaxExc("invalid char constant~");
  
    start = (int) token[0];  
  
    // discard closing '  
    gettoken(token);  
    if(token[0] != '\'')  
      throw SyntaxExc("' expected in cases.");  
  }  
  else  
   throw SyntaxExc("Constant expected in cases.");  
  
  skipspaces();  
  
  // Read and discard the "to".  
  gettoken(token);  
  if(token != "to")  
    throw SyntaxExc("to expected in cases.");  
  
  skipspaces();  
  
  // Get ending value.  
  gettoken(token);  
  
  if(isdigit(token[0])) {  
    // is an int constant  
    end = atoi(token.c_str());  
  }  
  else if(token[0] == '\'') {  
    // is char constant  
    gettoken(token);  
    if(token.size() != 1) throw SyntaxExc("invalid char constant~");
  
    end = (int) token[0];  
  
    // discard closing '  
    gettoken(token);  
    if(token[0] != '\'')  
      throw SyntaxExc("' expected in cases.");  
  }  
  else  
   throw SyntaxExc("Constant expected in cases.");  
  
  skipspaces();  
  
  // Read and discard the :  
  gettoken(token);  
  
  if(token != ":")   
    throw SyntaxExc(": expected in cases.");  
  
  // Generate stacked case statments.  
  fout << "case " << start << ":\n";  
  for(int i = start+1 ; i <= end; i++) {  
    fout << indent << "case " << i << ":";  
    if(i != end) fout << endl;  
  }  
}  
  
// Translate a repeat loop.  
void repeat() {  
  fout << "do";  
}    
  
// Translate an until.  
void until() {  
  string token;  
  int parencount = 1;  
  
  fout << "while";  
  
  skipspaces();  
  
  // Read and store the (  
  gettoken(token);  
  if(token != "(")  
    throw SyntaxExc("( expected in typeof.");  
  fout << "(";  
  
  // Begin while by reversing and  
  // parenthesizing the condition.  
  fout << "!(";  
  
  // Now, read the expression.  
  do {  
    if(!gettoken(token))  
      throw SyntaxExc("Unexpected EOF encountered.");  
  
    if(token == "(") parencount++;  
    if(token == ")") parencount--;  
  
    fout << token;  
  } while(parencount > 0);  
  fout << ")";  
}  
  
// Translate a typeof expression.   
void typeof() {  
  string token;  
  string temp;  
  
  fout << "typeid(";  
  
  skipspaces();  
  
  gettoken(token);  
  
  do {  
    temp = token;  
  
    if(!gettoken(token))  
      throw SyntaxExc("Unexpected EOF encountered.");  
    
    if(token != "same") fout << temp;  
  } while(token != "same");  
  
  skipspaces();  
  
  gettoken(token);  
  
  if(token != "as") throw SyntaxExc("as expected.");  
  
  fout << ") == typeid(";  
  
  skipspaces();  
  
  do {  
    if(!gettoken(token))  
      throw SyntaxExc("Unexpected EOF encountered.");  
  
    fout << token;  
  } while(token != ")");  
  fout << ")";  
}  
  
void skipspaces() {  
  char ch;  
  
  do {  
    ch = fin.get();  
  } while(isspace(ch));  
  fin.putback(ch);  
}


void breakon()
{
    fout << "if";
    string token;
    skipspaces();
    gettoken(token);

    if(token != "(")
        throw SyntaxExc("left bracket is required!");

    fout <<"(";
    skipspaces();
    
    int bracketCount = 1;

    do
    {
        gettoken(token);
        if(token == "(") bracketCount++;
        else if(token == ")") bracketCount--;
        fout << token;
    }while(bracketCount > 0);

    fout << "break";

    
}

void ignore()
{
    fout << "if";
    skipspaces();

    string token;
    gettoken(token);
    if(token != "(") 
        throw SyntaxExc("( is required here\n");

    fout << "(";
    int bracketCount = 1;

    do 
    {
        gettoken(token);
        if(token == "(") bracketCount++;
        else if(token == ")") bracketCount--;
        fout << token;
    } while (bracketCount > 0);

    fout << "continue";



}

void elif()
{
    fout << "else if";
}