beautify.cpp

Decompilation Dos Program is a technique that allows you to recover lost source code. It is also nee

#include "Disc.h"
#include "Beautify.h"
#include <string.h>
#include <stdio.h>
#include <conio.h>
#define IsHexDigit(x)  ((x>='0' && x<='9') || \
			(x>='A' && x<='F') || \
			(x>='a' && x<='f'))
#define IsAlphabet(x) ((x>='a' && x<='z') || (x>='A' && x<='Z') || x=='_')
/*---------------------------------------------------------------
DelRedundantParantheses - Removes a matching "()" if there are no
 operators inside the "()".
Called by DisplayProc only.
---------------------------------------------------------------*/
void CBeautifier::DelRedundantParantheses(char *tStr)
{
  static char OperChars[]="><=+-*/%!~^&|?,;";
  char *str = tStr;
  char *ptr;
  int CheckEntireString=(*tStr!='(');
  // Reach the 1st '('.
  while(*str && *str!='(') str++;
  // EndofString?
  if (!*str) return;
  // else...
  char *StringBegin = str;
  // Check till we reach the corresponding ')'
  do
  {
    str++;
    int OperatorFound=0;

    while(*str && *str!=')')
    {
      // Skip variables like [BP-02] etc.
      if (*str=='[') while(*str!=']') str++;
      // Another "()"? Check that by recursion.
      if (*str=='(')
      {
	// Is it a function call?
	if (!IsHexDigit(*(str-1)) && !IsAlphabet(*(str-1)))
	  DelRedundantParantheses(str);
      }
      // Still the '(' is present? If so, skip till the ')' is reached.
      if (*str=='(')
	while(*str!=')') str++;
      ptr = OperChars;
      while(*ptr) if (*str==*ptr++) break;
      // Did we find a operator match?
      if (*ptr) OperatorFound=1;
      // Check the next character.
      str++;
    }
    if (*str && !OperatorFound)
    {
      // Is it a function call?
      if (IsHexDigit(*(StringBegin-1))) return;
      // Is it any keyword followed by a '(' - like switch(x)?
      if (IsAlphabet(*(StringBegin-1))) return;
      // else...
      strcpy(StringBegin,StringBegin+1);
      str--;
      strcpy(str,str+1);
    }
  }while(*str && CheckEntireString);
}
/*
[-----------------------------------------------------------------------]
[ GetLabel - returns the address of the stringlist containing the given	]
[  string "Labelxxx".							]
[-----------------------------------------------------------------------]
*/
StringList *CBeautifier::GetLabel(StringList *sList,char *labelptr)
{
  int labellen = strlen(labelptr)-1;
  while(sList)
  {
    if (strncmp(sList->Data,labelptr,labellen)==0) return sList;
    sList = sList->Next;
  }
  return NULL;
}
/*----------------------------------------------------------------------
IsJumpingAcrossBlocks - checks if there are any new blocks created and
 not closed between srcList and tgtList.  Used while translating 'goto's
 to do-while loops and in translating 'if' conditions.
----------------------------------------------------------------------*/
int CBeautifier::IsJumpingAcrossBlocks(StringList *srcList,StringList *tgtList)
{
  srcList=srcList->Next;
  int NumUnmatchedBraces=0;
  int OnlyBracesFound=0;
  int i=srcList->GetPos(tgtList);
  while(i--)
  {
    if (srcList->Data[0]=='}')
    {
      NumUnmatchedBraces--;
      OnlyBracesFound=1;
    }
    else
    {
      OnlyBracesFound=0;
      if (srcList->Data[0]=='{') NumUnmatchedBraces++;
    }
    srcList = srcList->Next;
  }
  if (OnlyBracesFound) NumUnmatchedBraces=0;
  return (NumUnmatchedBraces!=0);
}
/*----------------------------------------------------------------------
BeautifyCode - Rearranges the 'if', 'goto' etc. in the form of loops.
----------------------------------------------------------------------*/
void CBeautifier::BeautifyCode(StringList *sList)
{
  StringList *ptr = sList;
  // start replacing if,goto statements with 'while' loops.
  ptr = sList;
  while(ptr)
  {
    if (!(ptr->Data)) break;
    // Conversion code for IF statement & DO-WHILE loop structures.
    if (strncmp(ptr->Data,"if",2)==0)
    {
      // check if the goto statement jumps to a line before the if stat.
      int PosOfIfGoto = sList->GetPos(GetLabel(sList,strstr(ptr->Data,"goto")+5));
      // if so,  this is not a conventional if..else statement.
      // it might be a DO-WHILE statement, checking is done below.
      if (PosOfIfGoto>sList->GetPos(ptr))
      {
	// Check if there is a JMP to from inside to outside a block.
	//  if so,  leave it unchanged.
	if (!IsJumpingAcrossBlocks(ptr,sList->GetList(PosOfIfGoto)))
	{
	  // Now, reverse the condition in the 'if' statement.
	  //  This is a bit difficult to understand.  The actual code is
	  //  generated to JMP if the condition is not satisfied.
	  //  Since we are transforming these conditions into 'if' blocks,
	  //  reversing the condition is necessary.
	  char *str = ptr->Data;
	  int i;
	  while(*str)
	  {
	    // 10 = Number of Jxx instructions decoded in the 'if' statement.
	    for(i=0;i<10;i++)
	      if (*str==CompOperators[i][0] && *(str+1)==CompOperators[i][1])
		break;
	    if (i!=10) break;
	    str++;
	  }
	  *str = OppCompOperators[i][0];
	  *(str+1) = OppCompOperators[i][1];

	  StringList *EndOfIf = sList->GetList(PosOfIfGoto-1);
	  // Is an else clause present?
	  if (strncmp(EndOfIf->Data,"goto",4)==0)
	  {
	    StringList *EndOfElse = GetLabel(sList,EndOfIf->Data+5);
	    // Is this position after the if clause?
	    if (sList->GetPos(EndOfElse)>sList->GetPos(ptr))
	    {
	      EndOfElse->Insert("}");
	      EndOfIf->Replace("}");
	      EndOfIf->Next->Insert("{");
	      EndOfIf->Next->Insert("else");
	    }
	    // else, there this is not the 'else' clause code.
	    else EndOfIf->Next->Insert("}");
	  }
	  else EndOfIf->Next->Insert("}");
	  *strstr(ptr->Data,"goto")=0;
	  ptr->Next->Insert("{");
	}
      }
      // check if it is a DO-WHILE loop structure.
      else
      {
	StringList *DoClause = sList->GetList(PosOfIfGoto);
	// Check if there is a JMP to from inside to outside a block(or back)
	//  if so,  leave it unchanged.
	if (!IsJumpingAcrossBlocks(DoClause,ptr))
	{
	  // Insert in the reverse order.
	  DoClause->Next->Insert("{");
	  DoClause->Next->Insert("do");

	  char whileclause[100];
	  strcpy(whileclause,"}while");
	  // copy the condition part of the if statement.
	  strcat(whileclause,ptr->Data+2);
	  // cut-off the goto part of the if statement.
	  *strstr(whileclause,"goto")=0;
	  strcat(whileclause,";");
	  ptr->Replace(whileclause);
	}
      }
    }
    // Conversion code for the WHILE loop structure.
    else if (strncmp(ptr->Data,"goto",4)==0)
    {
      StringList *PtrOfGotoLabel = GetLabel(ptr,ptr->Data+5);
      StringList *PtrOfIf = PtrOfGotoLabel->Next;
      if (PtrOfIf)
      {
	StringList *PtrOfIfGoto;
	// check if the goto statement jumps to a line after the if stat.
	int PosOfIfGoto = sList->GetPos(GetLabel(PtrOfIf,strstr(PtrOfIf->Data,"goto")+5));
	// If so,  this is not a while loop
	if (PosOfIfGoto<sList->GetPos(PtrOfIf))
	{
	  // check from before the if clause.
	  PtrOfIfGoto = GetLabel(ptr,strstr(PtrOfIf->Data,"goto")+5);
	  // found?
	  if (ptr->Next == PtrOfIfGoto)
	  {
	    char whileclause[100];
	    strcpy(whileclause,"while");
	    // copy the condition part of the if statement.
	    strcat(whileclause,PtrOfIf->Data+2);
	    // cut-off the goto part of the if statement.
	    *strstr(whileclause,"goto")=0;
	    ptr->Replace(whileclause);
	    ptr->Next->Insert("{");
	    PtrOfIf->Replace("}");
	  }
	}
      }
    }
    ptr = ptr->Next;
  }
  // Now do some beautification work.
  ptr = sList;
  while(ptr)
  {
    if (!ptr->Data) break;
    // trim down unwanted parantheses.
    DelRedundantParantheses(ptr->Data);
    ptr = ptr->Next;
  }
  DeleteUnusedLabels(sList);
}
void CBeautifier::DeleteUnusedLabels(StringList *sList)
{
  int NumUsedLabels;
  StringList *ptr;

  int NumLabelsfound=0;
  NumUsedLabels=0;
  int *LabelArr;
ComForFindAndStore:
  ptr = sList;
  while(ptr && ptr->Data)
  {
    char *str = strstr(ptr->Data,"goto");
    if (str)
    {
      if (NumLabelsfound)
      {
	int lab = atoi(str+10);
	for(int i=0;i<NumUsedLabels;i++)
	  if (LabelArr[i]==lab) break;
	if (i==NumUsedLabels) LabelArr[NumUsedLabels++] = lab;
      }
      else NumUsedLabels++;
    }
    ptr = ptr->Next;
  }
  if (!NumLabelsfound && NumUsedLabels)
  {
    LabelArr = new int[NumUsedLabels];
    assert(LabelArr);
    NumLabelsfound=1;
    NumUsedLabels=0;
    goto ComForFindAndStore;
  }
  ptr = sList;
  while(ptr && ptr->Data)
  {
    StringList *tmp = ptr;
    ptr = ptr->Next;
    if (strncmp(tmp->Data,"Label",5)==0)
    {
      int lab = atoi(tmp->Data+5);
      for(int i=0;i<NumUsedLabels;i++)
	if (LabelArr[i]==lab) break;
      if (i==NumUsedLabels)
	sList->Delete(tmp);
    }
  }
}
StringList *CBeautifier::DisplayList(StringList *sList,FILE *fp,int Indent,int InCase)
{
  while(sList && sList->Data)
  {
    if (sList->Data[0]=='}') return sList;
    else if (InCase)
    {
      if (strncmp(sList->Data,"default",6)==0 ||
	  strncmp(sList->Data,"case",4)==0) return sList;
    }
  PrintLine:

    if (sList->Data)
    {
      if ((int)sList->addr!=-1) fprintf(fp,"% 4X: ",sList->addr);
      else fprintf(fp,"      ");
      for(int i=0;i<Indent;i++) fprintf(fp," ");
      fprintf(fp,"%s\n",sList->Data);

      if (sList->Data[0]=='{')
      {
	sList = DisplayList(sList->Next,fp,Indent+2,0); goto PrintLine;
      }
      else if (strncmp(sList->Data,"case",4)==0)
      {
	sList = DisplayList(sList->Next,fp,Indent+2,1); goto PrintLine;
      }
      else if (strncmp(sList->Data,"default",7)==0)
      {
	sList = DisplayList(sList->Next,fp,Indent+2,1);
	return sList;
      }
      sList = sList->Next;
    }
  }
  return NULL;
}