/*
 * lineparse.cpp
 * 
 * This file is a part of NSIS.
 * 
 * Copyright (C) 1999-2020 Nullsoft and Contributors
 * 
 * Licensed under the zlib/libpng license (the "License");
 * you may not use this file except in compliance with the License.
 * 
 * Licence details can be found in the file COPYING.
 * 
 * This software is provided 'as-is', without any express or implied
 * warranty.
 */

#include "lineparse.h"
#include "Platform.h"
#include "tchar.h"
#include <cstdlib>
#include <cstring>
#include "tstring.h"
#include "util.h"

#ifndef _istspace
#define _istspace my_istspace
template<class T> static int my_istspace(T c) { return c == 0x20 || (c >= 0x09 && c <= 0x0D); }
#endif
template<class T> static T skipspace(T s) { while(*s && _istspace(*s)) ++s; return s; }


LineParser::LineParser(bool bCommentBlock)
{
  m_incommentblock=bCommentBlock;
  m_incomment=false;
  m_nt=m_eat=0;
  m_tokens=0;
}

LineParser::~LineParser()
{
  freetokens();
}

bool LineParser::inComment()
{
  return m_incomment;
}

bool LineParser::inCommentBlock()
{
  return m_incommentblock;
}

int LineParser::parse(TCHAR *line, int ignore_escaping/*=0*/) // returns -1 on error
{
  freetokens();
  bool bPrevCB=m_incommentblock;
  int n=doline(line, ignore_escaping);
  if (n) return n;
  if (m_nt)
  {
    m_incommentblock=bPrevCB;
    m_tokens=(TCHAR**)malloc(sizeof(TCHAR*)*m_nt);
    n=doline(line, ignore_escaping);
    if (n)
    {
      freetokens();
      return -1;
    }
  }
  return 0;
}

int LineParser::getnumtokens()
{
  return m_nt-m_eat;
}

void LineParser::eattoken()
{
  m_eat++;
}

inline int LineParser::validate_token_index(int token, int *success/*=0*/) const
{
  token += m_eat;
  if (token < 0 || token >= m_nt || !m_tokens[token][0])
  {
    if (success) *success = 0;
    return -1;
  }
  return token;
}


static unsigned int get_base_from_prefix(const TCHAR *str)
{
  unsigned int hasbase = 0;
  if (_T('-') == *str || _T('+') == *str) ++str;
  if (_T('0') == str[0])
  {
    if (_T('x') == (str[1]|32)) hasbase = 16;
    // Special support for 0n, 0y and 0t MASM style and 0b and 0o Python style radix prefix:
    if (_T('n') == (str[1]|32)) hasbase = 10;
    if (_T('b') == (str[1]|32) || _T('y') == (str[1]|32)) hasbase = 2;
    if (_T('o') == (str[1]|32) || _T('t') == (str[1]|32)) hasbase = 8;
  }
  return hasbase;
}
static unsigned int number_has_base_prefix(const TCHAR *str) { return get_base_from_prefix(str); }

static int parse_int_expression(const TCHAR *str, bool expressions, int *success=0)
{
  enum { XOP_NONE = 0, XOP_BITAND, XOP_BITXOR, XOP_BITOR, XOP_LOGAND, XOP_LOGOR };
  int tot=0, xop=XOP_NONE, hasfailed=0;
nextnum:
  const TCHAR *p=str, *parse=p, *xoperandpfixstart, *xoperandpfixend;
  TCHAR *end;
  int neg=0, base=0, num;
#ifndef _MSC_VER
  xoperandpfixstart=xoperandpfixend=0; // Avoid GCC maybe-uninitialized warning
#endif

  if (expressions)
  {
    xoperandpfixstart=p=skipspace(p);
    while (*p == '~' || *p == '!') p=skipspace(++p);
    xoperandpfixend=parse=p;
  }

  if (_T('+') == *p)
    ++p;
  else if (_T('-') == *p)
    ++p, ++neg;
  if (_T('0') == p[0])
  {
    if ((base=get_base_from_prefix(p))) parse=&p[2];
  }
  if (neg)
  {
    num=(int)(long)_tcstol(parse,&end,base);
    if (base) num*=-1; // Input was "-0n012" but we have only parsed "012" and need to fix the sign
  }
  else
  {
    num=(int)(long)_tcstoul(parse,&end,base);
  }
  const bool parseddigits=end > parse;
  if (!parseddigits) ++hasfailed; // "0x", "1 | ~" etc.

  if (expressions)
  {
    for (p=xoperandpfixstart; p < xoperandpfixend; p=skipspace(++p))
      switch(*p)
      {
      case '~': num=~num; break;
      case '!': num=!num; break;
      default: assert(!"op");
      }

    // Simple left-to-right parsing, no operator precedence
    switch(xop)
    {
    case XOP_NONE: tot=num; break;
    case XOP_BITAND: tot&=num; break;
    case XOP_BITXOR: tot^=num; break;
    case XOP_BITOR: tot|=num; break;
    case XOP_LOGAND: tot=tot && num; break;
    case XOP_LOGOR: tot=tot || num; break;
    default: assert(!"op");
    }

    str=end=skipspace(end);
    switch(*str)
    {
    case '&': xop=XOP_BITAND; if (*++str == '&') xop=XOP_LOGAND, ++str; goto nextnum;
    case '^': xop=XOP_BITXOR, ++str; goto nextnum;
    case '|': xop=XOP_BITOR; if (*++str == '|') xop=XOP_LOGOR, ++str; goto nextnum;
    default: xop=XOP_NONE; // Done
    }
  }
  else
  {
    tot=num;
  }

  if (success) *success=!(int)(*end) && !hasfailed;
  return tot;
}

int LineParser::parse_int(const TCHAR *str, int *success/*=0*/)
{
  return parse_int_expression(str, false, success);
}

int LineParser::parse_intx(const TCHAR *str, int *success/*=0*/)
{
  return parse_int_expression(str, true, success);
}

int LineParser::gettoken_int(int token, int *success/*=0*/) const
{
  if ((token = validate_token_index(token,success)) < 0) return 0;
  return parse_int(m_tokens[token],success);
}

int LineParser::gettoken_intx(int token, int *success/*=0*/) const
{
  if ((token = validate_token_index(token,success)) < 0) return 0;
  return parse_intx(m_tokens[token],success);
}

double LineParser::parse_float(const TCHAR *str, int *success/*=0*/)
{
  if (success)
  {
    const TCHAR *t=str;
    if (_T('-') == *t || _T('+') == *t) ++t;
    *success=*t ? 1 : 0;
    unsigned int dotcount = 0;
    while (*t)
    {
      if (_T('.') == *t && ++dotcount > 1) *success=0;
      if ((*t < _T('0') || *t > _T('9')) && *t != _T('.')) *success=0;
      t++;
    }
  }
  return _tstof(str);
}

double LineParser::gettoken_float(int token, int *success/*=0*/) const
{
  if ((token = validate_token_index(token,success)) < 0) return 0.0;
  return parse_float(m_tokens[token],success);
}

double LineParser::parse_number(const TCHAR *str, int *success/*=0*/)
{
  const unsigned int forceint=number_has_base_prefix(str);
  return forceint ? parse_int(str,success) : parse_float(str,success);
}

double LineParser::gettoken_number(int token, int *success/*=0*/) const
{
  if ((token = validate_token_index(token,success)) < 0) return 0.0;
  return parse_number(m_tokens[token],success);
}

int LineParser::gettoken_binstrdata(int token, char*buffer, int bufcap) const
{
  const TCHAR*p=gettoken_str(token);
  int a,b,c,d=0;
  while (*p)
  {
    a=*p;
    if (a >= _T('0') && a <= _T('9')) a-=_T('0');
    else if (a >= _T('a') && a <= _T('f')) a-=_T('a')-10;
    else if (a >= _T('A') && a <= _T('F')) a-=_T('A')-10;
    else if (a == _T(',')) { ++p; continue; } // Allow comma separator (for Regedit5 .reg format)
    else break;
    b=*++p;
    if (b >= _T('0') && b <= _T('9')) b-=_T('0');
    else if (b >= _T('a') && b <= _T('f')) b-=_T('a')-10;
    else if (b >= _T('A') && b <= _T('F')) b-=_T('A')-10;
    else break;
    c=(a<<4)|b, p++;
    if (d >= bufcap) return -1; // Buffer too small
    buffer[d++]=c;
  }
  if (*p) return -2; // Did not parse the entire buffer
  return d;
}

TCHAR* LineParser::gettoken_str(int token) const
{
  token+=m_eat;
  if (token < 0 || token >= m_nt) return const_cast<TCHAR*>(_T(""));
  return m_tokens[token];
}

int LineParser::gettoken_enum(int token, const TCHAR *strlist) // null separated list
{
  int x=0;
  TCHAR *tt=gettoken_str(token);
  if (tt && *tt) while (*strlist)
  {
    if (!_tcsicmp(tt,strlist)) return x;
    strlist+=_tcslen(strlist)+1;
    x++;
  }
  return -1;
}

void LineParser::freetokens()
{
  if (m_tokens)
  {
    int x;
    for (x = 0; x < m_nt; x ++)
      free(m_tokens[x]);
    free(m_tokens);
  }
  m_tokens=0;
  m_nt=0;
}

int LineParser::doline(TCHAR *line, int ignore_escaping/*=0*/)
{
  m_nt=0;
  m_incomment = false;
  while (*line == _T(' ') || *line == _T('\t')) line++;
  while (*line)
  {
    if ( m_incommentblock )
    {
      while ( *line )
      {
        if ( *line == _T('*') && *(line+1) == _T('/') )
        {
          m_incommentblock=false; // Found end of comment block
          line+=2;
          while (*line == _T(' ') || *line == _T('\t')) line++;
          break;
        }
        else line++;
      }
    }
    else
    {
      int lstate=0; // 1=", 2=`, 4='
      if (*line == _T(';') || *line == _T('#'))
      {
        m_incomment = true;
        break;
      }
      if (*line == _T('/') && *(line+1) == _T('*'))
      {
        m_incommentblock = true;
        line+=2;
      }
      else
      {
        if (*line == _T('\"')) lstate=1;
        else if (*line == _T('\'')) lstate=2;
        else if (*line == _T('`')) lstate=4;
        if (lstate) line++;
        int nc=0;
        TCHAR *p = line;
        while (*line)
        {
          if (line[0] == _T('$') && line[1] == _T('\\'))
          {
            switch (line[2]) 
            {
              case _T('"'):
              case _T('\''):
              case _T('`'):
                nc += ignore_escaping ? 3 : 1;
                line += 3;
                continue;
            }
          }
          if (lstate==1 && *line ==_T('\"')) break;
          if (lstate==2 && *line ==_T('\'')) break;
          if (lstate==4 && *line ==_T('`')) break;
          if (!lstate && (*line == _T(' ') || *line == _T('\t'))) break;
#ifdef NSIS_FIX_COMMENT_HANDLING
          if (!lstate && (*line == _T(';') || *line == _T('#') || (*line == _T('/') && *(line+1) == _T('*')))) break;
#endif
          line++;
          nc++;
        }
        if (m_tokens)
        {
          int i;
          m_tokens[m_nt]=(TCHAR*)malloc((nc+1)*sizeof(TCHAR));
          for (i = 0; p < line; i++, p++) {
            if (!ignore_escaping && p[0] == _T('$') && p[1] == _T('\\'))
            {
              switch (p[2])
              {
                case _T('"'):
                case _T('\''):
                case _T('`'):
                  p += 2;
              }
            }
            m_tokens[m_nt][i] = *p;
          }
          m_tokens[m_nt][nc]=0;
        }
        m_nt++;
        if (lstate)
        {
          if (*line) line++;
          else return -2;
        }
        while (*line == _T(' ') || *line == _T('\t')) line++;
      }
    }
  }
  return 0;
}