cppcheck/tokenize.cpp

1426 lines
41 KiB
C++
Raw Normal View History

2008-11-20 20:23:05 +01:00
/*
2008-10-26 08:55:15 +01:00
* c++check - c/c++ syntax checking
* Copyright (C) 2007 Daniel Marjamäki
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/
*/
2007-05-24 07:40:45 +02:00
//---------------------------------------------------------------------------
#include "tokenize.h"
2007-05-24 07:40:45 +02:00
//---------------------------------------------------------------------------
#include <locale>
#include <fstream>
#include <string>
#include <cstring>
#include <iostream>
#include <sstream>
#include <list>
#include <algorithm>
2007-05-24 07:40:45 +02:00
#include <stdlib.h> // <- strtoul
2008-12-14 23:03:49 +01:00
#include <stdio.h>
#include <filelister.h>
2007-05-24 07:40:45 +02:00
2008-02-18 18:11:34 +01:00
#ifdef __BORLANDC__
#include <ctype.h>
2008-02-18 18:11:34 +01:00
#include <mem.h>
#endif
#ifndef _MSC_VER
#define _strdup(str) strdup(str)
#endif
2007-05-24 07:40:45 +02:00
//---------------------------------------------------------------------------
Tokenizer::Tokenizer()
{
_tokens = 0;
_tokensBack = 0;
_dsymlist = 0;
}
Tokenizer::~Tokenizer()
{
DeallocateTokens();
}
//---------------------------------------------------------------------------
// Helper functions..
TOKEN *Tokenizer::_gettok(TOKEN *tok, int index)
{
while (tok && index>0)
{
tok = tok->next();
index--;
}
return tok;
2008-12-09 01:49:32 +01:00
}
//---------------------------------------------------------------------------
const TOKEN *Tokenizer::tokens() const
{
return _tokens;
}
//---------------------------------------------------------------------------
// Defined symbols.
// "#define abc 123" will create a defined symbol "abc" with the value 123
//---------------------------------------------------------------------------
const std::vector<std::string> *Tokenizer::getFiles() const
{
return &_files;
}
2007-05-24 07:40:45 +02:00
void Tokenizer::Define(const char Name[], const char Value[])
2007-05-24 07:40:45 +02:00
{
if (!(Name && Name[0]))
return;
if (!(Value && Value[0]))
return;
// Is 'Value' a decimal value..
bool dec = true, hex = true;
for (int i = 0; Value[i]; i++)
{
if ( ! isdigit(Value[i]) )
2007-05-24 07:40:45 +02:00
dec = false;
if ( ! isxdigit(Value[i]) && (!(i==1 && Value[i]=='x')))
2007-05-24 07:40:45 +02:00
hex = false;
}
if (!dec && !hex)
return;
char *strValue = _strdup(Value);
2007-05-24 07:40:45 +02:00
if (!dec && hex)
{
// Convert Value from hexadecimal to decimal
unsigned long value;
std::istringstream istr(Value+2);
istr >> std::hex >> value;
std::ostringstream ostr;
ostr << value;
2007-05-24 07:40:45 +02:00
free(strValue);
strValue = _strdup(ostr.str().c_str());
2007-05-24 07:40:45 +02:00
}
DefineSymbol *NewSym = new DefineSymbol;
memset(NewSym, 0, sizeof(DefineSymbol));
NewSym->name = _strdup(Name);
2007-05-24 07:40:45 +02:00
NewSym->value = strValue;
NewSym->next = _dsymlist;
_dsymlist = NewSym;
2007-05-24 07:40:45 +02:00
}
//---------------------------------------------------------------------------
//---------------------------------------------------------------------------
// addtoken
// add a token. Used by 'Tokenizer'
//---------------------------------------------------------------------------
2007-05-24 07:40:45 +02:00
void Tokenizer::addtoken(const char str[], const unsigned int lineno, const unsigned int fileno)
2007-05-24 07:40:45 +02:00
{
if (str[0] == 0)
return;
// Replace hexadecimal value with decimal
std::ostringstream str2;
if (strncmp(str,"0x",2)==0)
2007-05-24 07:40:45 +02:00
{
str2 << strtoul(str+2, NULL, 16);
2007-05-24 07:40:45 +02:00
}
else
{
str2 << str;
}
2007-05-24 07:40:45 +02:00
if (_tokensBack)
2008-12-13 12:39:36 +01:00
{
_tokensBack->insertToken( str2.str().c_str() );
_tokensBack = _tokensBack->next();
2007-05-24 07:40:45 +02:00
}
else
{
2008-12-13 12:39:36 +01:00
_tokens = new TOKEN;
_tokensBack = _tokens;
_tokensBack->setstr( str2.str().c_str() );
2008-12-13 12:39:36 +01:00
}
_tokensBack->linenr( lineno );
_tokensBack->fileIndex( fileno );
2007-05-24 07:40:45 +02:00
// Check if str is defined..
for (DefineSymbol *sym = _dsymlist; sym; sym = sym->next)
2007-05-24 07:40:45 +02:00
{
if (strcmp(str,sym->name)==0)
{
_tokensBack->setstr(sym->value);
2007-05-24 07:40:45 +02:00
break;
}
}
}
//---------------------------------------------------------------------------
//---------------------------------------------------------------------------
// SizeOfType - gives the size of a type
//---------------------------------------------------------------------------
int Tokenizer::SizeOfType(const char type[]) const
{
if (!type)
return 0;
std::map<std::string, unsigned int>::const_iterator it = _typeSize.find(type);
if ( it == _typeSize.end() )
return 0;
return it->second;
2007-05-24 07:40:45 +02:00
}
//---------------------------------------------------------------------------
2007-05-29 08:24:36 +02:00
//---------------------------------------------------------------------------
// InsertTokens - Copy and insert tokens
//---------------------------------------------------------------------------
void Tokenizer::InsertTokens(TOKEN *dest, TOKEN *src, unsigned int n)
2007-05-29 08:24:36 +02:00
{
while (n > 0)
2008-12-13 12:39:36 +01:00
{
dest->insertToken( src->aaaa() );
dest->next()->fileIndex( src->fileIndex() );
dest->next()->linenr( src->linenr() );
dest = dest->next();
src = src->next();
2007-05-29 08:24:36 +02:00
n--;
}
}
//---------------------------------------------------------------------------
//---------------------------------------------------------------------------
// Tokenize - tokenizes a given file.
//---------------------------------------------------------------------------
void Tokenizer::tokenize(std::istream &code, const char FileName[])
2007-05-24 07:40:45 +02:00
{
// Has this file been tokenized already?
for (unsigned int i = 0; i < _files.size(); i++)
2007-05-24 07:40:45 +02:00
{
if ( SameFileName( _files[i].c_str(), FileName ) )
2007-05-24 07:40:45 +02:00
return;
}
// The "_files" vector remembers what files have been tokenized..
_files.push_back( FileLister::simplifyPath( FileName ) );
2007-05-24 07:40:45 +02:00
2008-02-16 16:46:32 +01:00
// Tokenize the file..
tokenizeCode( code, (unsigned int)(_files.size() - 1) );
2008-02-16 16:46:32 +01:00
}
//---------------------------------------------------------------------------
//---------------------------------------------------------------------------
// Tokenize - tokenizes input stream
//---------------------------------------------------------------------------
void Tokenizer::tokenizeCode(std::istream &code, const unsigned int FileIndex)
2008-02-16 16:46:32 +01:00
{
// Tokenize the file.
2007-05-24 07:40:45 +02:00
unsigned int lineno = 1;
std::string CurrentToken;
for (char ch = (char)code.get(); code.good(); ch = (char)code.get())
2007-05-24 07:40:45 +02:00
{
// We are not handling UTF and stuff like that. Code is supposed to plain simple text.
if ( ch < 0 )
continue;
// Preprocessor stuff?
if (ch == '#' && CurrentToken.empty())
2007-05-24 07:40:45 +02:00
{
std::string line("#");
{
char chPrev = '#';
while ( code.good() )
{
ch = (char)code.get();
if (chPrev!='\\' && ch=='\n')
break;
if (ch!=' ')
chPrev = ch;
if (ch!='\\' && ch!='\n')
line += ch;
if (ch=='\n')
++lineno;
}
}
2007-05-24 07:40:45 +02:00
if (strncmp(line.c_str(),"#include",8)==0 &&
line.find("\"") != std::string::npos)
{
// Extract the filename
line.erase(0, line.find("\"")+1);
line.erase(line.find("\""));
// Relative path..
if (_files.back().find_first_of("\\/") != std::string::npos)
2007-05-24 07:40:45 +02:00
{
std::string path = _files.back();
2008-02-16 16:46:32 +01:00
path.erase( 1 + path.find_last_of("\\/") );
2007-05-24 07:40:45 +02:00
line = path + line;
}
2008-02-16 16:46:32 +01:00
addtoken("#include", lineno, FileIndex);
addtoken(line.c_str(), lineno, FileIndex);
2007-05-24 07:40:45 +02:00
std::ifstream fin( line.c_str() );
tokenize(fin, line.c_str());
2007-05-24 07:40:45 +02:00
}
else if (strncmp(line.c_str(), "#define", 7) == 0)
{
std::string strId;
2007-05-24 07:40:45 +02:00
enum {Space1, Id, Space2, Value} State;
State = Space1;
for (unsigned int i = 8; i < line.length(); i++)
{
if (State==Space1 || State==Space2)
{
if (isspace(line[i]))
2007-05-24 07:40:45 +02:00
continue;
State = (State==Space1) ? Id : Value;
}
else if (State==Id)
2007-05-24 07:40:45 +02:00
{
if ( isspace( line[i] ) )
{
strId = CurrentToken;
CurrentToken.clear();
State = Space2;
continue;
}
else if ( ! isalnum(line[i]) )
{
break;
}
2007-05-24 07:40:45 +02:00
}
CurrentToken += line[i];
2007-05-24 07:40:45 +02:00
}
if (State==Value)
{
2008-02-16 16:46:32 +01:00
addtoken("def", lineno, FileIndex);
addtoken(strId.c_str(), lineno, FileIndex);
2008-02-16 16:46:32 +01:00
addtoken(";", lineno, FileIndex);
Define(strId.c_str(), CurrentToken.c_str());
2007-05-24 07:40:45 +02:00
}
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
}
else
{
2008-02-16 16:46:32 +01:00
addtoken("#", lineno, FileIndex);
addtoken(";", lineno, FileIndex);
}
2007-05-24 07:40:45 +02:00
lineno++;
continue;
}
if (ch == '\n')
{
// Add current token..
addtoken(CurrentToken.c_str(), lineno++, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
continue;
}
// Comments..
if (ch == '/' && code.good())
2007-05-24 07:40:45 +02:00
{
bool newstatement = bool( strchr(";{}", CurrentToken.empty() ? '\0' : CurrentToken[0]) != NULL );
2007-05-24 07:40:45 +02:00
// Add current token..
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
// Read next character..
2008-02-16 16:46:32 +01:00
ch = (char)code.get();
2007-05-24 07:40:45 +02:00
// If '//'..
if (ch == '/')
{
std::string comment;
getline( code, comment ); // Parse in the whole comment
// If the comment says something like "fred is deleted" then generate appropriate tokens for that
comment = comment + " ";
if ( newstatement && comment.find(" deleted ")!=std::string::npos )
{
// delete
addtoken( "delete", lineno, FileIndex );
// fred
std::string::size_type pos1 = comment.find_first_not_of(" \t");
std::string::size_type pos2 = comment.find(" ", pos1);
std::string firstWord = comment.substr( pos1, pos2-pos1 );
addtoken( firstWord.c_str(), lineno, FileIndex );
// ;
addtoken( ";", lineno, FileIndex );
}
2007-05-24 07:40:45 +02:00
lineno++;
continue;
}
// If '/*'..
if (ch == '*')
{
char chPrev;
ch = chPrev = 'A';
while (code.good() && (chPrev!='*' || ch!='/'))
2007-05-24 07:40:45 +02:00
{
chPrev = ch;
2008-02-16 16:46:32 +01:00
ch = (char)code.get();
2007-05-24 07:40:45 +02:00
if (ch == '\n')
lineno++;
}
continue;
}
// Not a comment.. add token..
2008-02-16 16:46:32 +01:00
addtoken("/", lineno, FileIndex);
2007-05-24 07:40:45 +02:00
}
// char..
if (ch == '\'')
{
// Add previous token
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
// Read this ..
CurrentToken += ch;
CurrentToken += (char)code.get();
CurrentToken += (char)code.get();
2007-05-24 07:40:45 +02:00
if (CurrentToken[1] == '\\')
CurrentToken += (char)code.get();
2007-05-24 07:40:45 +02:00
// Add token and start on next..
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
continue;
}
// String..
if (ch == '\"')
{
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
bool special = false;
char c = ch;
do
{
// Append token..
2008-12-09 01:49:32 +01:00
CurrentToken += c;
if ( c == '\n' )
++lineno;
2007-05-24 07:40:45 +02:00
// Special sequence '\.'
if (special)
special = false;
else
special = (c == '\\');
// Get next character
2008-02-16 16:46:32 +01:00
c = (char)code.get();
2007-05-24 07:40:45 +02:00
}
while (code.good() && (special || c != '\"'));
CurrentToken += '\"';
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
continue;
}
if (strchr("+-*/%&|^?!=<>[](){};:,.~",ch))
2007-05-24 07:40:45 +02:00
{
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
CurrentToken += ch;
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
continue;
}
if (isspace(ch) || iscntrl(ch))
2007-05-24 07:40:45 +02:00
{
addtoken(CurrentToken.c_str(), lineno, FileIndex);
CurrentToken.clear();
2007-05-24 07:40:45 +02:00
continue;
}
CurrentToken += ch;
2007-05-24 07:40:45 +02:00
}
addtoken( CurrentToken.c_str(), lineno, FileIndex );
2007-05-24 07:40:45 +02:00
// Combine tokens..
for (TOKEN *tok = _tokens; tok && tok->next(); tok = tok->next())
2007-05-24 07:40:45 +02:00
{
tok->combineWithNext("<", "<");
tok->combineWithNext(">", ">");
tok->combineWithNext("&", "&");
tok->combineWithNext("|", "|");
tok->combineWithNext("+", "=");
tok->combineWithNext("-", "=");
tok->combineWithNext("*", "=");
tok->combineWithNext("/", "=");
tok->combineWithNext("&", "=");
tok->combineWithNext("|", "=");
tok->combineWithNext("=", "=");
tok->combineWithNext("!", "=");
tok->combineWithNext("<", "=");
tok->combineWithNext(">", "=");
tok->combineWithNext(":", ":");
tok->combineWithNext("-", ">");
tok->combineWithNext("private", ":");
tok->combineWithNext("protected", ":");
tok->combineWithNext("public", ":");
2007-05-24 07:40:45 +02:00
}
// Replace "->" with "."
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( tok->str() == "->" )
{
tok->setstr(".");
}
}
2008-08-28 08:37:11 +02:00
// typedef..
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
2008-08-28 08:37:11 +02:00
{
if (TOKEN::Match(tok, "typedef %type% %type% ;"))
2008-08-28 08:37:11 +02:00
{
const char *type1 = tok->strAt( 1);
const char *type2 = tok->strAt( 2);
for ( TOKEN *tok2 = tok; tok2; tok2 = tok2->next() )
2008-08-28 08:37:11 +02:00
{
if (tok2->aaaa()!=type1 && tok2->aaaa()!=type2 && (tok2->str() == type2))
2008-08-28 08:37:11 +02:00
{
tok2->setstr(type1);
2008-08-28 08:37:11 +02:00
}
}
}
else if (TOKEN::Match(tok, "typedef %type% %type% %type% ;"))
2008-08-28 08:37:11 +02:00
{
const char *type1 = tok->strAt( 1);
const char *type2 = tok->strAt( 2);
const char *type3 = tok->strAt( 3);
TOKEN *tok2 = tok;
while ( ! TOKEN::Match(tok2, ";") )
tok2 = tok2->next();
for ( ; tok2; tok2 = tok2->next() )
2008-08-28 08:37:11 +02:00
{
if (tok2->aaaa()!=type3 && (tok2->str() == type3))
2008-08-28 08:37:11 +02:00
{
tok2->setstr(type1);
2008-12-13 12:39:36 +01:00
tok2->insertToken( type2 );
tok2->next()->fileIndex( tok2->fileIndex() );
tok2->next()->linenr( tok2->linenr() );
tok2 = tok2->next();
2008-08-28 08:37:11 +02:00
}
}
}
}
// Remove __asm..
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( TOKEN::Match(tok->next(), "__asm {") )
{
while ( tok->next() )
{
bool last = TOKEN::Match( tok->next(), "}" );
2008-12-13 12:39:36 +01:00
// Unlink and delete tok->next()
tok->eraseToken();
// break if this was the last token to delete..
if (last)
break;
}
}
2008-12-09 01:49:32 +01:00
}
// Remove "volatile"
while ( TOKEN::Match(_tokens, "volatile") )
{
TOKEN *tok = _tokens;
_tokens = _tokens->next();
delete tok;
}
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
while ( TOKEN::Match(tok->next(), "volatile") )
{
tok->deleteNext();
}
}
2007-05-29 08:24:36 +02:00
}
//---------------------------------------------------------------------------
2008-12-09 01:49:32 +01:00
void Tokenizer::setVarId()
{
// Clear all variable ids
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
tok->varId( 0 );
// Set variable ids..
unsigned int _varId = 0;
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( ! TOKEN::Match(tok, "[;{}(] %type% %var%") )
continue;
// Determine name of declared variable..
const char *varname = 0;
TOKEN *tok2 = tok->next();
while ( tok2 && ! TOKEN::Match( tok2, "[;[=(]" ) )
2008-12-09 01:49:32 +01:00
{
if ( tok2->isName() )
varname = tok2->strAt(0);
else if ( tok2->str() != "*" )
break;
tok2 = tok2->next();
}
// Variable declaration found => Set variable ids
if ( TOKEN::Match(tok2, "[;[=]") && varname )
{
++_varId;
int indentlevel = 0;
int parlevel = 0;
for ( tok2 = tok->next(); tok2 && indentlevel >= 0; tok2 = tok2->next() )
{
if ( tok2->str() == varname )
tok2->varId( _varId );
else if ( tok2->str() == "{" )
++indentlevel;
else if ( tok2->str() == "}" )
--indentlevel;
else if ( tok2->str() == "(" )
++parlevel;
else if ( tok2->str() == ")" )
--parlevel;
else if ( parlevel < 0 && tok2->str() == ";" )
break;
}
}
}
}
2007-05-29 08:24:36 +02:00
//---------------------------------------------------------------------------
// Simplify token list
//---------------------------------------------------------------------------
void Tokenizer::simplifyTokenList()
2007-05-29 08:24:36 +02:00
{
// Remove the keyword 'unsigned'
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if (tok->next() && (tok->next()->str() == "unsigned"))
2008-12-09 01:49:32 +01:00
{
tok->deleteNext();
}
}
2007-05-24 07:40:45 +02:00
// Replace constants..
for (TOKEN *tok = _tokens; tok; tok = tok->next())
2007-05-24 07:40:45 +02:00
{
if (TOKEN::Match(tok,"const %type% %var% = %num% ;"))
2007-05-24 07:40:45 +02:00
{
const char *sym = tok->strAt(2);
const char *num = tok->strAt(4);
for (TOKEN *tok2 = _gettok(tok,6); tok2; tok2 = tok2->next())
2007-05-24 07:40:45 +02:00
{
if (tok2->str() == sym)
2007-05-24 07:40:45 +02:00
{
tok2->setstr(num);
2007-05-24 07:40:45 +02:00
}
}
}
}
// Fill the map _typeSize..
_typeSize.clear();
_typeSize["char"] = sizeof(char);
_typeSize["short"] = sizeof(short);
_typeSize["int"] = sizeof(int);
_typeSize["long"] = sizeof(long);
_typeSize["float"] = sizeof(float);
_typeSize["double"] = sizeof(double);
for (TOKEN *tok = _tokens; tok; tok = tok->next())
{
if (TOKEN::Match(tok,"class %var%"))
{
_typeSize[tok->strAt(1)] = 11;
}
else if (TOKEN::Match(tok, "struct %var%"))
{
_typeSize[tok->strAt(1)] = 13;
}
}
// Replace 'sizeof(type)'..
for (TOKEN *tok = _tokens; tok; tok = tok->next())
{
if (tok->str() != "sizeof")
continue;
if (TOKEN::Match(tok, "sizeof ( %type% * )"))
{
std::ostringstream str;
// 'sizeof(type *)' has the same size as 'sizeof(char *)'
str << sizeof(char *);
tok->setstr( str.str().c_str() );
2007-05-24 07:40:45 +02:00
for (int i = 0; i < 4; i++)
{
tok->deleteNext();
}
}
else if (TOKEN::Match(tok, "sizeof ( %type% )"))
{
const char *type = tok->strAt( 2);
int size = SizeOfType(type);
if (size > 0)
{
std::ostringstream str;
str << size;
tok->setstr( str.str().c_str() );
for (int i = 0; i < 3; i++)
{
tok->deleteNext();
}
}
}
else if (TOKEN::Match(tok, "sizeof ( * %var% )"))
{
tok->setstr("100");
for ( int i = 0; i < 4; ++i )
tok->deleteNext();
}
2007-05-24 07:40:45 +02:00
}
// Replace 'sizeof(var)'
for (TOKEN *tok = _tokens; tok; tok = tok->next())
{
// type array [ num ] ;
if ( ! TOKEN::Match(tok, "%type% %var% [ %num% ] ;") )
continue;
int size = SizeOfType(tok->aaaa());
if (size <= 0)
continue;
const char *varname = tok->strAt( 1);
int total_size = size * atoi( tok->strAt( 3) );
// Replace 'sizeof(var)' with number
int indentlevel = 0;
for ( TOKEN *tok2 = _gettok(tok,5); tok2; tok2 = tok2->next() )
{
if (tok2->str() == "{")
{
++indentlevel;
}
else if (tok2->str() == "}")
{
--indentlevel;
if (indentlevel < 0)
break;
}
// Todo: TOKEN::Match varname directly
else if (TOKEN::Match(tok2, "sizeof ( %var% )"))
{
if (strcmp(tok2->strAt(2), varname) == 0)
{
std::ostringstream str;
str << total_size;
tok2->setstr(str.str().c_str());
// Delete the other tokens..
for (int i = 0; i < 3; i++)
{
tok2->deleteNext();
}
}
}
}
}
2007-05-29 08:24:36 +02:00
// Simple calculations..
for ( bool done = false; !done; done = true )
{
for (TOKEN *tok = _tokens; tok; tok = tok->next())
{
if (TOKEN::Match(tok->next(), "* 1") || TOKEN::Match(tok->next(), "1 *"))
{
for (int i = 0; i < 2; i++)
tok->deleteNext();
done = false;
}
2008-12-09 01:49:32 +01:00
// (1-2)
if (TOKEN::Match(tok, "[[,(=<>] %num% [+-*/] %num% [],);=<>]"))
{
int i1 = atoi(tok->strAt(1));
int i2 = atoi(tok->strAt(3));
if ( i2 == 0 && *(tok->strAt(2)) == '/' )
{
continue;
}
switch (*(tok->strAt(2)))
{
case '+': i1 += i2; break;
case '-': i1 -= i2; break;
case '*': i1 *= i2; break;
case '/': i1 /= i2; break;
}
tok = tok->next();
std::ostringstream str;
str << i1;
tok->setstr(str.str().c_str());
for (int i = 0; i < 2; i++)
{
tok->deleteNext();
}
2008-03-22 12:46:06 +01:00
done = false;
}
}
}
// Replace "*(str + num)" => "str[num]"
for (TOKEN *tok = _tokens; tok; tok = tok->next())
{
if ( ! strchr(";{}(=<>", tok->aaaa0()) )
continue;
TOKEN *next = tok->next();
if ( ! next )
break;
if (TOKEN::Match(next, "* ( %var% + %num% )"))
{
const char *str[4] = {"var","[","num","]"};
str[0] = tok->strAt(3);
str[2] = tok->strAt(5);
for (int i = 0; i < 4; i++)
{
tok = tok->next();
tok->setstr(str[i]);
}
tok->deleteNext();
tok->deleteNext();
}
}
2007-05-29 08:24:36 +02:00
// Split up variable declarations if possible..
for (TOKEN *tok = _tokens; tok; tok = tok->next())
2007-05-29 08:24:36 +02:00
{
if ( ! TOKEN::Match(tok, "[{};]") )
2007-05-29 08:24:36 +02:00
continue;
TOKEN *type0 = tok->next();
if (!TOKEN::Match(type0, "%type%"))
continue;
if (TOKEN::Match(type0, "else") || TOKEN::Match(type0, "return"))
continue;
2007-05-29 08:24:36 +02:00
TOKEN *tok2 = NULL;
unsigned int typelen = 0;
if ( TOKEN::Match(type0, "%type% %var% ,|=") )
2008-12-09 01:49:32 +01:00
{
if ( type0->next()->str() != "operator" )
{
tok2 = _gettok(type0, 2); // The ',' or '=' token
2008-12-09 01:49:32 +01:00
typelen = 1;
}
2007-05-29 08:24:36 +02:00
}
else if ( TOKEN::Match(type0, "%type% * %var% ,|=") )
2008-12-09 01:49:32 +01:00
{
if ( type0->next()->next()->str() != "operator" )
{
tok2 = _gettok(type0, 3); // The ',' token
2008-12-09 01:49:32 +01:00
typelen = 1;
}
2007-05-29 08:24:36 +02:00
}
else if ( TOKEN::Match(type0, "%type% %var% [ %num% ] ,|=") )
2007-05-29 08:24:36 +02:00
{
2008-03-22 12:46:06 +01:00
tok2 = _gettok(type0, 5); // The ',' token
2007-05-29 08:24:36 +02:00
typelen = 1;
}
else if ( TOKEN::Match(type0, "%type% * %var% [ %num% ] ,|=") )
2007-05-29 08:24:36 +02:00
{
2008-03-22 12:46:06 +01:00
tok2 = _gettok(type0, 6); // The ',' token
2007-05-29 19:11:53 +02:00
typelen = 1;
}
else if ( TOKEN::Match(type0, "struct %type% %var% ,|=") )
2007-05-29 19:11:53 +02:00
{
2008-03-22 12:46:06 +01:00
tok2 = _gettok(type0, 3);
2007-05-29 19:11:53 +02:00
typelen = 2;
}
else if ( TOKEN::Match(type0, "struct %type% * %var% ,|=") )
2007-05-29 19:11:53 +02:00
{
2008-03-22 12:46:06 +01:00
tok2 = _gettok(type0, 4);
2007-05-29 08:24:36 +02:00
typelen = 2;
}
if (tok2)
{
2008-12-05 19:25:08 +01:00
if (tok2->str() == ",")
2007-05-29 08:24:36 +02:00
{
tok2->setstr(";");
2007-05-29 08:24:36 +02:00
InsertTokens(tok2, type0, typelen);
}
else
{
TOKEN *eq = tok2;
int parlevel = 0;
while (tok2)
{
if ( strchr("{(", tok2->aaaa0()) )
2007-05-29 08:24:36 +02:00
{
parlevel++;
}
else if ( strchr("})", tok2->aaaa0()) )
2007-05-29 08:24:36 +02:00
{
if (parlevel<0)
break;
parlevel--;
}
else if ( parlevel==0 && strchr(";,",tok2->aaaa0()) )
2007-05-29 08:24:36 +02:00
{
// "type var =" => "type var; var ="
2008-03-22 12:46:06 +01:00
TOKEN *VarTok = _gettok(type0,typelen);
if (VarTok->aaaa0()=='*')
VarTok = VarTok->next();
2007-05-29 19:11:53 +02:00
InsertTokens(eq, VarTok, 2);
eq->setstr(";");
2007-05-29 08:24:36 +02:00
// "= x, " => "= x; type "
if (tok2->str() == ",")
2007-05-29 08:24:36 +02:00
{
tok2->setstr(";");
2007-05-29 08:24:36 +02:00
InsertTokens( tok2, type0, typelen );
}
break;
}
tok2 = tok2->next();
2007-05-29 08:24:36 +02:00
}
}
}
}
// Replace NULL with 0..
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( TOKEN::Match(tok, "NULL") )
tok->setstr("0");
}
// Replace pointer casts of 0.. "(char *)0" => "0"
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( TOKEN::Match(tok->next(), "( %type% * ) 0") || TOKEN::Match(tok->next(),"( %type% %type% * ) 0") )
{
while (!TOKEN::Match(tok->next(),"0"))
tok->deleteNext();
}
}
2008-12-14 18:05:21 +01:00
bool done = false;
while ( ! done )
2008-12-14 18:05:21 +01:00
{
done = true;
2008-12-14 18:05:21 +01:00
if( simplifyConditions() )
done = false;
if( simplifyCasts() )
done = false;
if( simplifyFunctionReturn() )
done = false;
}
2007-05-24 07:40:45 +02:00
}
//---------------------------------------------------------------------------
bool Tokenizer::simplifyConditions()
{
bool ret = false;
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if (TOKEN::Match(tok, "( true &&") || TOKEN::Match(tok, "&& true &&") || TOKEN::Match(tok->next(), "&& true )"))
{
tok->deleteNext();
tok->deleteNext();
ret = true;
}
else if (TOKEN::Match(tok, "( false ||") || TOKEN::Match(tok, "|| false ||") || TOKEN::Match(tok->next(), "|| false )"))
{
tok->deleteNext();
tok->deleteNext();
ret = true;
}
// Change numeric constant in condition to "true" or "false"
const TOKEN *tok2 = tok->tokAt(2);
if ((TOKEN::Match(tok, "(") || TOKEN::Match(tok, "&&") || TOKEN::Match(tok, "||")) &&
TOKEN::Match(tok->next(), "%num%") &&
(TOKEN::Match(tok2, ")") || TOKEN::Match(tok2, "&&") || TOKEN::Match(tok2, "||")) )
{
tok->next()->setstr((tok->next()->str() != "0") ? "true" : "false");
ret = true;
2008-12-09 01:49:32 +01:00
}
// Reduce "(%num% == %num%)" => "(true)"/"(false)"
if ( (TOKEN::Match(tok, "&&") || TOKEN::Match(tok, "||") || TOKEN::Match(tok, "(")) &&
TOKEN::Match(tok->tokAt(1), "%num% %any% %num%") &&
(TOKEN::Match(tok->tokAt(4), "&&") || TOKEN::Match(tok->tokAt(4), "||") || TOKEN::Match(tok->tokAt(4), ")")) )
{
double op1 = (strstr(tok->strAt(1), "0x")) ? strtol(tok->strAt(1),0,16) : atof( tok->strAt(1) );
double op2 = (strstr(tok->strAt(3), "0x")) ? strtol(tok->strAt(3),0,16) : atof( tok->strAt(3) );
std::string cmp = tok->strAt(2);
bool result = false;
if ( cmp == "==" )
result = (op1 == op2);
else if ( cmp == "!=" )
result = (op1 != op2);
else if ( cmp == ">=" )
result = (op1 >= op2);
else if ( cmp == ">" )
result = (op1 > op2);
else if ( cmp == "<=" )
result = (op1 <= op2);
else if ( cmp == "<" )
result = (op1 < op2);
else
cmp = "";
if ( ! cmp.empty() )
{
tok = tok->next();
tok->deleteNext();
tok->deleteNext();
tok->setstr( result ? "true" : "false" );
ret = true;
2008-12-09 01:49:32 +01:00
}
}
}
return ret;
}
2008-12-14 18:05:21 +01:00
bool Tokenizer::simplifyCasts()
{
bool ret = false;
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( TOKEN::Match(tok->next(), "( %type% * )") )
{
tok->deleteNext();
tok->deleteNext();
tok->deleteNext();
tok->deleteNext();
ret = true;
}
else if ( TOKEN::Match(tok->next(), "dynamic_cast|reinterpret_cast|const_cast|static_cast <" ) )
{
while ( tok->next() && tok->next()->str() != ">" )
tok->deleteNext();
tok->deleteNext();
tok->deleteNext();
TOKEN *tok2 = tok;
int parlevel = 0;
while ( tok2->next() && parlevel >= 0 )
{
tok2 = tok2->next();
if ( TOKEN::Match(tok2->next(), "(") )
++parlevel;
else if ( TOKEN::Match(tok2->next(), ")") )
--parlevel;
}
if (tok2->next())
tok2->deleteNext();
ret = true;
}
}
return ret;
}
bool Tokenizer::simplifyFunctionReturn()
{
bool ret = false;
int indentlevel = 0;
for ( const TOKEN *tok = tokens(); tok; tok = tok->next() )
{
if ( tok->str() == "{" )
++indentlevel;
else if ( tok->str() == "}" )
--indentlevel;
else if ( indentlevel == 0 && TOKEN::Match(tok, "%var% ( ) { return %num% ; }") )
{
std::ostringstream pattern;
pattern << "[(=+-*/] " << tok->str() << " ( ) [;)+-*/]";
for ( TOKEN *tok2 = _tokens; tok2; tok2 = tok2->next() )
{
if ( TOKEN::Match(tok2, pattern.str().c_str()) )
{
tok2 = tok2->next();
tok2->setstr( tok->strAt(5) );
tok2->deleteNext();
tok2->deleteNext();
ret = true;
}
}
}
}
return ret;
}
bool Tokenizer::simplifyKnownVariables()
{
// TODO, this function should be called from simplifyTokenList()
// after the implementation is done.
// TODO, this functions needs to be implemented.
// TODO, test
bool ret = false;
for ( TOKEN *tok = _tokens; tok; tok = tok->next() )
{
// Search for a block of code
if ( ! TOKEN::Match(tok, ") const| {") )
continue;
// parse the block of code..
int indentlevel = 0;
for ( TOKEN *tok2 = tok; tok2; tok2 = tok2->next() )
{
if ( tok2->str() == "{" )
++indentlevel;
else if ( tok2->str() == "}" )
{
--indentlevel;
if ( indentlevel <= 0 )
continue;
}
else if ( TOKEN::Match(tok2, "%var% = %num% ;") )
{
unsigned int varid = tok2->varId();
TOKEN *tok3 = tok2;
while ( tok3 )
{
tok3 = tok3->next();
// Perhaps it's a loop => bail out
if ( TOKEN::Match(tok3, "[{}]") )
break;
// Variable is used somehow in a non-defined pattern => bail out
if ( tok3->varId() == varid )
break;
// Replace variable with numeric constant..
if ( TOKEN::Match(tok3, "if ( %varid% )", 0, 0, varid) )
{
tok3 = tok3->next()->next();
tok3->setstr( tok2->strAt(2) );
ret = true;
}
}
}
}
}
return ret;
}
//---------------------------------------------------------------------------
// Helper functions for handling the tokens list
//---------------------------------------------------------------------------
2007-05-29 19:11:53 +02:00
2007-05-29 08:24:36 +02:00
//---------------------------------------------------------------------------
const TOKEN *Tokenizer::GetFunctionTokenByName( const char funcname[] ) const
{
for ( unsigned int i = 0; i < _functionList.size(); ++i )
{
if ( _functionList[i]->str() == funcname )
{
return _functionList[i];
}
}
return NULL;
}
void Tokenizer::fillFunctionList()
{
_functionList.clear();
bool staticfunc = false;
bool classfunc = false;
int indentlevel = 0;
for ( const TOKEN *tok = _tokens; tok; tok = tok->next() )
{
if ( tok->str() == "{" )
++indentlevel;
else if ( tok->str() == "}" )
--indentlevel;
if (indentlevel > 0)
{
continue;
}
if (strchr("};", tok->aaaa0()))
staticfunc = classfunc = false;
else if ( tok->str() == "static" )
staticfunc = true;
else if ( tok->str() == "::" )
classfunc = true;
else if (TOKEN::Match(tok, "%var% ("))
{
// Check if this is the first token of a function implementation..
for ( const TOKEN *tok2 = tok; tok2; tok2 = tok2->next() )
{
if ( tok2->str() == ";" )
{
tok = tok2;
break;
}
else if ( tok2->str() == "{" )
{
break;
}
else if ( tok2->str() == ")" )
2008-12-09 01:49:32 +01:00
{
if ( TOKEN::Match(tok2, ") const| {") )
{
_functionList.push_back( tok );
tok = tok2;
}
else
{
tok = tok2;
while (tok->next() && !strchr(";{", tok->next()->aaaa0()))
tok = tok->next();
}
break;
}
}
}
}
// If the _functionList functions with duplicate names, remove them
// TODO this will need some better handling
for ( unsigned int func1 = 0; func1 < _functionList.size(); )
2008-12-09 01:49:32 +01:00
{
bool hasDuplicates = false;
for ( unsigned int func2 = func1 + 1; func2 < _functionList.size(); )
{
if ( _functionList[func1]->str() == _functionList[func2]->str() )
{
hasDuplicates = true;
_functionList.erase( _functionList.begin() + func2 );
}
else
{
++func2;
}
}
if ( ! hasDuplicates )
{
++func1;
}
else
{
_functionList.erase( _functionList.begin() + func1 );
}
2008-12-09 01:49:32 +01:00
}
}
//---------------------------------------------------------------------------
2007-05-29 08:24:36 +02:00
// Deallocate lists..
void Tokenizer::DeallocateTokens()
{
deleteTokens( _tokens );
_tokens = 0;
_tokensBack = 0;
while (_dsymlist)
{
struct DefineSymbol *next = _dsymlist->next;
free(_dsymlist->name);
free(_dsymlist->value);
delete _dsymlist;
_dsymlist = next;
}
_files.clear();
}
void Tokenizer::deleteTokens(TOKEN *tok)
{
while (tok)
{
TOKEN *next = tok->next();
delete tok;
tok = next;
}
2008-12-09 01:49:32 +01:00
}
//---------------------------------------------------------------------------
const char *Tokenizer::getParameterName( const TOKEN *ftok, int par )
{
int _par = 1;
for ( ; ftok; ftok = ftok->next())
{
if ( TOKEN::Match(ftok, ",") )
++_par;
if ( par==_par && TOKEN::Match(ftok, "%var% [,)]") )
return ftok->aaaa();
}
return NULL;
}
//---------------------------------------------------------------------------
std::string Tokenizer::fileLine( const TOKEN *tok ) const
{
std::ostringstream ostr;
ostr << "[" << _files.at(tok->fileIndex()) << ":" << tok->linenr() << "]";
return ostr.str();
}
//---------------------------------------------------------------------------
bool Tokenizer::SameFileName( const char fname1[], const char fname2[] )
{
#ifdef __linux__
return bool( strcmp(fname1, fname2) == 0 );
#endif
#ifdef __GNUC__
return bool( strcasecmp(fname1, fname2) == 0 );
#endif
#ifdef __BORLANDC__
return bool( stricmp(fname1, fname2) == 0 );
#endif
#ifdef _MSC_VER
return bool( _stricmp(fname1, fname2) == 0 );
#endif
}
//---------------------------------------------------------------------------