cppcheck/lib/tokenize.cpp

11147 lines
411 KiB
C++

/*
* Cppcheck - A tool for static C/C++ code analysis
* Copyright (C) 2007-2019 Cppcheck team.
*
* 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/>.
*/
//---------------------------------------------------------------------------
#include "tokenize.h"
#include "check.h"
#include "library.h"
#include "mathlib.h"
#include "path.h"
#include "platform.h"
#include "settings.h"
#include "standards.h"
#include "symboldatabase.h"
#include "templatesimplifier.h"
#include "timer.h"
#include "token.h"
#include "utils.h"
#include "valueflow.h"
#include <algorithm>
#include <cassert>
#include <cctype>
#include <cstring>
#include <ctime>
#include <iostream>
#include <stack>
#include <unordered_map>
#include <utility>
#include <vector>
//---------------------------------------------------------------------------
namespace {
// local struct used in setVarId
// in order to store information about the scope
struct VarIdScopeInfo {
VarIdScopeInfo()
:isExecutable(false), isStructInit(false), isEnum(false), startVarid(0) {
}
VarIdScopeInfo(bool _isExecutable, bool _isStructInit, bool _isEnum, unsigned int _startVarid)
:isExecutable(_isExecutable), isStructInit(_isStructInit), isEnum(_isEnum), startVarid(_startVarid) {
}
const bool isExecutable;
const bool isStructInit;
const bool isEnum;
const unsigned int startVarid;
};
}
/** Return whether tok is the "{" that starts an enumerator list */
static bool isEnumStart(const Token* tok)
{
if (!tok || tok->str() != "{")
return false;
return (tok->strAt(-1) == "enum") || (tok->strAt(-2) == "enum");
}
template<typename T>
static void skipEnumBody(T **tok)
{
T *defStart = *tok;
while (Token::Match(defStart, "%name%|::|:"))
defStart = defStart->next();
if (defStart && defStart->str() == "{")
*tok = defStart->link()->next();
}
const Token * Tokenizer::isFunctionHead(const Token *tok, const std::string &endsWith) const
{
return Tokenizer::isFunctionHead(tok, endsWith, isCPP());
}
const Token * Tokenizer::isFunctionHead(const Token *tok, const std::string &endsWith, bool cpp)
{
if (!tok)
return nullptr;
if (tok->str() == "(")
tok = tok->link();
if (Token::Match(tok, ") ;|{|[")) {
tok = tok->next();
while (tok && tok->str() == "[" && tok->link())
tok = tok->link()->next();
return (tok && endsWith.find(tok->str()) != std::string::npos) ? tok : nullptr;
}
if (cpp && tok->str() == ")") {
tok = tok->next();
while (Token::Match(tok, "const|noexcept|override|final|volatile|&|&& !!(") ||
(Token::Match(tok, "%name% !!(") && tok->isUpperCaseName()))
tok = tok->next();
if (tok && tok->str() == ")")
tok = tok->next();
while (tok && tok->str() == "[")
tok = tok->link()->next();
if (Token::Match(tok, "throw|noexcept ("))
tok = tok->linkAt(1)->next();
if (Token::Match(tok, "%name% (") && tok->isUpperCaseName())
tok = tok->linkAt(1)->next();
if (tok && tok->str() == ".") { // trailing return type
for (tok = tok->next(); tok && !Token::Match(tok, ";|{|override|final"); tok = tok->next())
if (tok->link() && Token::Match(tok, "<|[|("))
tok = tok->link();
}
while (Token::Match(tok, "override|final !!(") ||
(Token::Match(tok, "%name% !!(") && tok->isUpperCaseName()))
tok = tok->next();
if (Token::Match(tok, "= 0|default|delete ;"))
tok = tok->tokAt(2);
return (tok && endsWith.find(tok->str()) != std::string::npos) ? tok : nullptr;
}
return nullptr;
}
/**
* is tok the start brace { of a class, struct, union, or enum
*/
static bool isClassStructUnionEnumStart(const Token * tok)
{
if (!Token::Match(tok->previous(), "class|struct|union|enum|%name%|>|>> {"))
return false;
const Token * tok2 = tok->previous();
while (tok2 && !Token::Match(tok2, "class|struct|union|enum|{|}|;"))
tok2 = tok2->previous();
return Token::Match(tok2, "class|struct|union|enum");
}
//---------------------------------------------------------------------------
Tokenizer::Tokenizer() :
list(nullptr),
mSettings(nullptr),
mErrorLogger(nullptr),
mSymbolDatabase(nullptr),
mTemplateSimplifier(nullptr),
mVarId(0),
mUnnamedCount(0),
mCodeWithTemplates(false), //is there any templates?
mTimerResults(nullptr)
#ifdef MAXTIME
,mMaxTime(std::time(0) + MAXTIME)
#endif
{
}
Tokenizer::Tokenizer(const Settings *settings, ErrorLogger *errorLogger) :
list(settings),
mSettings(settings),
mErrorLogger(errorLogger),
mSymbolDatabase(nullptr),
mTemplateSimplifier(nullptr),
mVarId(0),
mUnnamedCount(0),
mCodeWithTemplates(false), //is there any templates?
mTimerResults(nullptr)
#ifdef MAXTIME
,mMaxTime(std::time(0) + MAXTIME)
#endif
{
// make sure settings are specified
assert(mSettings);
mTemplateSimplifier = new TemplateSimplifier(this);
}
Tokenizer::~Tokenizer()
{
delete mSymbolDatabase;
delete mTemplateSimplifier;
}
//---------------------------------------------------------------------------
// SizeOfType - gives the size of a type
//---------------------------------------------------------------------------
unsigned int Tokenizer::sizeOfType(const Token *type) const
{
if (!type || type->str().empty())
return 0;
if (type->tokType() == Token::eString)
return Token::getStrLength(type) + 1U;
const std::map<std::string, unsigned int>::const_iterator it = mTypeSize.find(type->str());
if (it == mTypeSize.end()) {
const Library::PodType* podtype = mSettings->library.podtype(type->str());
if (!podtype)
return 0;
return podtype->size;
} else if (type->isLong()) {
if (type->str() == "double")
return mSettings->sizeof_long_double;
else if (type->str() == "long")
return mSettings->sizeof_long_long;
}
return it->second;
}
//---------------------------------------------------------------------------
// check if this statement is a duplicate definition
bool Tokenizer::duplicateTypedef(Token **tokPtr, const Token *name, const Token *typeDef) const
{
// check for an end of definition
const Token * tok = *tokPtr;
if (tok && Token::Match(tok->next(), ";|,|[|=|)|>|(|{")) {
const Token * end = tok->next();
if (end->str() == "[") {
if (!end->link())
syntaxError(end); // invalid code
end = end->link()->next();
} else if (end->str() == ",") {
// check for derived class
if (Token::Match(tok->previous(), "public|private|protected"))
return false;
// find end of definition
while (end && end->next() && !Token::Match(end->next(), ";|)|>")) {
if (end->next()->str() == "(")
end = end->linkAt(1);
end = (end)?end->next():nullptr;
}
if (end)
end = end->next();
} else if (end->str() == "(") {
if (tok->previous()->str().compare(0, 8, "operator") == 0) {
// conversion operator
return false;
} else if (tok->previous()->str() == "typedef") {
// typedef of function returning this type
return false;
} else if (Token::Match(tok->previous(), "public:|private:|protected:")) {
return false;
} else if (tok->previous()->str() == ">") {
if (!Token::Match(tok->tokAt(-2), "%type%"))
return false;
if (!Token::Match(tok->tokAt(-3), ",|<"))
return false;
*tokPtr = end->link();
return true;
}
}
if (end) {
if (Token::simpleMatch(end, ") {")) { // function parameter ?
// look backwards
if (Token::Match(tok->previous(), "%type%") &&
!Token::Match(tok->previous(), "return|new|const|struct")) {
// duplicate definition so skip entire function
*tokPtr = end->next()->link();
return true;
}
} else if (end->str() == ">") { // template parameter ?
// look backwards
if (Token::Match(tok->previous(), "%type%") &&
!Token::Match(tok->previous(), "return|new|const|volatile")) {
// duplicate definition so skip entire template
while (end && end->str() != "{")
end = end->next();
if (end) {
*tokPtr = end->link();
return true;
}
}
} else {
// look backwards
if (Token::Match(tok->previous(), "typedef|}|>") ||
(end->str() == ";" && tok->previous()->str() == ",") ||
(tok->previous()->str() == "*" && tok->next()->str() != "(") ||
(Token::Match(tok->previous(), "%type%") &&
(!Token::Match(tok->previous(), "return|new|const|friend|public|private|protected|throw|extern") &&
!Token::simpleMatch(tok->tokAt(-2), "friend class")))) {
// scan backwards for the end of the previous statement
while (tok && tok->previous() && !Token::Match(tok->previous(), ";|{")) {
if (tok->previous()->str() == "}") {
tok = tok->previous()->link();
} else if (tok->previous()->str() == "typedef") {
return true;
} else if (tok->previous()->str() == "enum") {
return true;
} else if (tok->previous()->str() == "struct") {
if (tok->strAt(-2) == "typedef" &&
tok->next()->str() == "{" &&
typeDef->strAt(3) != "{") {
// declaration after forward declaration
return true;
} else if (tok->next()->str() == "{") {
return true;
} else if (Token::Match(tok->next(), ")|*")) {
return true;
} else if (tok->next()->str() == name->str()) {
return true;
} else if (tok->next()->str() != ";") {
return true;
} else {
return false;
}
} else if (tok->previous()->str() == "union") {
if (tok->next()->str() != ";") {
return true;
} else {
return false;
}
} else if (isCPP() && tok->previous()->str() == "class") {
if (tok->next()->str() != ";") {
return true;
} else {
return false;
}
}
if (tok)
tok = tok->previous();
}
if ((*tokPtr)->strAt(1) != "(" || !Token::Match((*tokPtr)->linkAt(1), ") .|(|["))
return true;
}
}
}
}
return false;
}
void Tokenizer::unsupportedTypedef(const Token *tok) const
{
if (!mSettings->debugwarnings)
return;
std::ostringstream str;
const Token *tok1 = tok;
unsigned int level = 0;
while (tok) {
if (level == 0 && tok->str() == ";")
break;
else if (tok->str() == "{")
++level;
else if (tok->str() == "}") {
if (level == 0)
break;
--level;
}
if (tok != tok1)
str << " ";
str << tok->str();
tok = tok->next();
}
if (tok)
str << " ;";
reportError(tok1, Severity::debug, "debug",
"Failed to parse \'" + str.str() + "\'. The checking continues anyway.");
}
Token * Tokenizer::deleteInvalidTypedef(Token *typeDef)
{
Token *tok = nullptr;
// remove typedef but leave ;
while (typeDef->next()) {
if (typeDef->next()->str() == ";") {
typeDef->deleteNext();
break;
} else if (typeDef->next()->str() == "{")
Token::eraseTokens(typeDef, typeDef->linkAt(1));
else if (typeDef->next()->str() == "}")
break;
typeDef->deleteNext();
}
if (typeDef != list.front()) {
tok = typeDef->previous();
tok->deleteNext();
} else {
list.front()->deleteThis();
tok = list.front();
}
return tok;
}
namespace {
struct Space {
Space() : bodyEnd(nullptr), isNamespace(false) { }
std::string className;
const Token * bodyEnd;
bool isNamespace;
};
}
static Token *splitDefinitionFromTypedef(Token *tok, unsigned int *unnamedCount)
{
Token *tok1;
std::string name;
bool isConst = false;
if (tok->next()->str() == "const") {
tok->deleteNext();
isConst = true;
}
if (tok->strAt(2) == "{") { // unnamed
tok1 = tok->linkAt(2);
if (tok1 && tok1->next()) {
// use typedef name if available
if (Token::Match(tok1->next(), "%type%"))
name = tok1->next()->str();
else // create a unique name
name = "Unnamed" + MathLib::toString((*unnamedCount)++);
tok->next()->insertToken(name);
} else
return nullptr;
} else if (tok->strAt(3) == ":") {
tok1 = tok->tokAt(4);
while (tok1 && tok1->str() != "{")
tok1 = tok1->next();
if (!tok1)
return nullptr;
tok1 = tok1->link();
name = tok->strAt(2);
} else { // has a name
tok1 = tok->linkAt(3);
if (!tok1)
return nullptr;
name = tok->strAt(2);
}
tok1->insertToken(";");
tok1 = tok1->next();
if (tok1->next() && tok1->next()->str() == ";" && tok1->previous()->str() == "}") {
tok->deleteThis();
tok1->deleteThis();
return nullptr;
} else {
tok1->insertToken("typedef");
tok1 = tok1->next();
Token * tok3 = tok1;
if (isConst) {
tok1->insertToken("const");
tok1 = tok1->next();
}
tok1->insertToken(tok->next()->str()); // struct, union or enum
tok1 = tok1->next();
tok1->insertToken(name);
tok->deleteThis();
tok = tok3;
}
return tok;
}
/* This function is called when processing function related typedefs.
* If simplifyTypedef generates an "Internal Error" message and the
* code that generated it deals in some way with functions, then this
* function will probably need to be extended to handle a new function
* related pattern */
Token *Tokenizer::processFunc(Token *tok2, bool inOperator) const
{
if (tok2->next() && tok2->next()->str() != ")" &&
tok2->next()->str() != ",") {
// skip over tokens for some types of canonicalization
if (Token::Match(tok2->next(), "( * %type% ) ("))
tok2 = tok2->linkAt(5);
else if (Token::Match(tok2->next(), "* ( * %type% ) ("))
tok2 = tok2->linkAt(6);
else if (Token::Match(tok2->next(), "* ( * %type% ) ;"))
tok2 = tok2->tokAt(5);
else if (Token::Match(tok2->next(), "* ( %type% [") &&
Token::Match(tok2->linkAt(4), "] ) ;|="))
tok2 = tok2->linkAt(4)->next();
else if (Token::Match(tok2->next(), "* ( * %type% ("))
tok2 = tok2->linkAt(5)->next();
else if (Token::simpleMatch(tok2->next(), "* [") &&
Token::simpleMatch(tok2->linkAt(2), "] ;"))
tok2 = tok2->next();
else {
if (tok2->next()->str() == "(")
tok2 = tok2->next()->link();
else if (!inOperator && !Token::Match(tok2->next(), "[|>|;")) {
tok2 = tok2->next();
while (Token::Match(tok2, "*|&") &&
!Token::Match(tok2->next(), ")|>"))
tok2 = tok2->next();
// skip over namespace
while (Token::Match(tok2, "%name% ::"))
tok2 = tok2->tokAt(2);
if (!tok2)
return nullptr;
if (tok2->str() == "(" &&
tok2->link()->next() &&
tok2->link()->next()->str() == "(") {
tok2 = tok2->link();
if (tok2->next()->str() == "(")
tok2 = tok2->next()->link();
}
// skip over typedef parameter
if (tok2->next() && tok2->next()->str() == "(") {
tok2 = tok2->next()->link();
if (!tok2->next())
syntaxError(tok2);
if (tok2->next()->str() == "(")
tok2 = tok2->next()->link();
}
}
}
}
return tok2;
}
void Tokenizer::simplifyTypedef()
{
std::vector<Space> spaceInfo;
bool isNamespace = false;
std::string className;
bool hasClass = false;
bool goback = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (mErrorLogger && !list.getFiles().empty())
mErrorLogger->reportProgress(list.getFiles()[0], "Tokenize (typedef)", tok->progressValue());
if (mSettings->terminated())
return;
if (isMaxTime())
return;
if (goback) {
//jump back once, see the comment at the end of the function
goback = false;
tok = tok->previous();
}
if (tok->str() != "typedef") {
if (tok->str() == "(" && tok->strAt(1) == "typedef") {
// Skip typedefs inside parentheses (#2453 and #4002)
tok = tok->next();
} else if (Token::Match(tok, "class|struct|namespace %any%") &&
(!tok->previous() || tok->previous()->str() != "enum")) {
isNamespace = (tok->str() == "namespace");
hasClass = true;
className = tok->next()->str();
} else if (hasClass && tok->str() == ";") {
hasClass = false;
} else if (hasClass && tok->str() == "{") {
Space info;
info.isNamespace = isNamespace;
info.className = className;
info.bodyEnd = tok->link();
spaceInfo.push_back(info);
hasClass = false;
} else if (!spaceInfo.empty() && tok->str() == "}" && spaceInfo.back().bodyEnd == tok) {
spaceInfo.pop_back();
}
continue;
}
// pull struct, union, enum or class definition out of typedef
// use typedef name for unnamed struct, union, enum or class
if (Token::Match(tok->next(), "const| struct|enum|union|class %type%| {")) {
Token *tok1 = splitDefinitionFromTypedef(tok, &mUnnamedCount);
if (!tok1)
continue;
tok = tok1;
} else if (Token::Match(tok->next(), "const| struct|class %type% :")) {
Token *tok1 = tok;
while (tok1 && tok1->str() != ";" && tok1->str() != "{")
tok1 = tok1->next();
if (tok1 && tok1->str() == "{") {
tok1 = splitDefinitionFromTypedef(tok, &mUnnamedCount);
if (!tok1)
continue;
tok = tok1;
}
}
/** @todo add support for union */
if (Token::Match(tok->next(), "enum %type% %type% ;") && tok->strAt(2) == tok->strAt(3)) {
tok->deleteNext(3);
tok->deleteThis();
if (tok->next())
tok->deleteThis();
//now the next token to process is 'tok', not 'tok->next()';
goback = true;
continue;
}
Token *typeName;
Token *typeStart = nullptr;
Token *typeEnd = nullptr;
Token *argStart = nullptr;
Token *argEnd = nullptr;
Token *arrayStart = nullptr;
Token *arrayEnd = nullptr;
Token *specStart = nullptr;
Token *specEnd = nullptr;
Token *typeDef = tok;
Token *argFuncRetStart = nullptr;
Token *argFuncRetEnd = nullptr;
Token *funcStart = nullptr;
Token *funcEnd = nullptr;
Token *tokOffset = tok->next();
bool function = false;
bool functionPtr = false;
bool functionRetFuncPtr = false;
bool functionPtrRetFuncPtr = false;
bool ptrToArray = false;
bool refToArray = false;
bool ptrMember = false;
bool typeOf = false;
Token *namespaceStart = nullptr;
Token *namespaceEnd = nullptr;
// check for invalid input
if (!tokOffset)
syntaxError(tok);
if (tokOffset->str() == "::") {
typeStart = tokOffset;
tokOffset = tokOffset->next();
while (Token::Match(tokOffset, "%type% ::"))
tokOffset = tokOffset->tokAt(2);
typeEnd = tokOffset;
if (Token::Match(tokOffset, "%type%"))
tokOffset = tokOffset->next();
} else if (Token::Match(tokOffset, "%type% ::")) {
typeStart = tokOffset;
do {
tokOffset = tokOffset->tokAt(2);
} while (Token::Match(tokOffset, "%type% ::"));
typeEnd = tokOffset;
if (Token::Match(tokOffset, "%type%"))
tokOffset = tokOffset->next();
} else if (Token::Match(tokOffset, "%type%")) {
typeStart = tokOffset;
while (Token::Match(tokOffset, "const|struct|enum %type%") ||
(tokOffset->next() && tokOffset->next()->isStandardType()))
tokOffset = tokOffset->next();
typeEnd = tokOffset;
tokOffset = tokOffset->next();
while (Token::Match(tokOffset, "%type%") &&
(tokOffset->isStandardType() || Token::Match(tokOffset, "unsigned|signed"))) {
typeEnd = tokOffset;
tokOffset = tokOffset->next();
}
bool atEnd = false;
while (!atEnd) {
if (tokOffset && tokOffset->str() == "::") {
typeEnd = tokOffset;
tokOffset = tokOffset->next();
}
if (Token::Match(tokOffset, "%type%") &&
tokOffset->next() && !Token::Match(tokOffset->next(), "[|;|,|(")) {
typeEnd = tokOffset;
tokOffset = tokOffset->next();
} else if (Token::simpleMatch(tokOffset, "const (")) {
typeEnd = tokOffset;
tokOffset = tokOffset->next();
atEnd = true;
} else
atEnd = true;
}
} else
continue; // invalid input
// check for invalid input
if (!tokOffset)
syntaxError(tok);
// check for template
if (!isC() && tokOffset->str() == "<") {
typeEnd = tokOffset->findClosingBracket();
while (typeEnd && Token::Match(typeEnd->next(), ":: %type%"))
typeEnd = typeEnd->tokAt(2);
if (!typeEnd) {
// internal error
return;
}
while (Token::Match(typeEnd->next(), "const|volatile"))
typeEnd = typeEnd->next();
tok = typeEnd;
tokOffset = tok->next();
}
std::list<std::string> pointers;
// check for pointers and references
while (Token::Match(tokOffset, "*|&|&&|const")) {
pointers.push_back(tokOffset->str());
tokOffset = tokOffset->next();
}
// check for invalid input
if (!tokOffset)
syntaxError(tok);
if (Token::Match(tokOffset, "%type%")) {
// found the type name
typeName = tokOffset;
tokOffset = tokOffset->next();
// check for array
if (tokOffset && tokOffset->str() == "[") {
arrayStart = tokOffset;
bool atEnd = false;
while (!atEnd) {
while (tokOffset->next() && !Token::Match(tokOffset->next(), ";|,")) {
tokOffset = tokOffset->next();
}
if (!tokOffset->next())
return; // invalid input
else if (tokOffset->next()->str() == ";")
atEnd = true;
else if (tokOffset->str() == "]")
atEnd = true;
else
tokOffset = tokOffset->next();
}
arrayEnd = tokOffset;
tokOffset = tokOffset->next();
}
// check for end or another
if (Token::Match(tokOffset, ";|,"))
tok = tokOffset;
// or a function typedef
else if (tokOffset && tokOffset->str() == "(") {
Token *tokOffset2 = nullptr;
if (Token::Match(tokOffset, "( *|%name%")) {
tokOffset2 = tokOffset->next();
if (tokOffset2->str() == "typename")
tokOffset2 = tokOffset2->next();
while (Token::Match(tokOffset2, "%type% ::"))
tokOffset2 = tokOffset2->tokAt(2);
}
// unhandled typedef, skip it and continue
if (typeName->str() == "void") {
unsupportedTypedef(typeDef);
tok = deleteInvalidTypedef(typeDef);
if (tok == list.front())
//now the next token to process is 'tok', not 'tok->next()';
goback = true;
continue;
}
// function pointer
else if (Token::Match(tokOffset2, "* %name% ) (")) {
// name token wasn't a name, it was part of the type
typeEnd = typeEnd->next();
functionPtr = true;
funcStart = funcEnd = tokOffset2; // *
tokOffset = tokOffset2->tokAt(3); // (
typeName = tokOffset->tokAt(-2);
argStart = tokOffset;
argEnd = tokOffset->link();
tok = argEnd->next();
}
// function
else if (isFunctionHead(tokOffset->link(), ";,")) {
function = true;
if (tokOffset->link()->next()->str() == "const") {
specStart = tokOffset->link()->next();
specEnd = specStart;
}
argStart = tokOffset;
argEnd = tokOffset->link();
tok = argEnd->next();
if (specStart)
tok = tok->next();
}
// syntax error
else
syntaxError(tok);
}
// unhandled typedef, skip it and continue
else {
unsupportedTypedef(typeDef);
tok = deleteInvalidTypedef(typeDef);
if (tok == list.front())
//now the next token to process is 'tok', not 'tok->next()';
goback = true;
continue;
}
}
// typeof: typedef __typeof__ ( ... ) type;
else if (Token::simpleMatch(tokOffset->previous(), "__typeof__ (") &&
Token::Match(tokOffset->link(), ") %type% ;")) {
argStart = tokOffset;
argEnd = tokOffset->link();
typeName = tokOffset->link()->next();
tok = typeName->next();
typeOf = true;
}
// function: typedef ... ( ... type )( ... );
// typedef ... (( ... type )( ... ));
// typedef ... ( * ( ... type )( ... ));
else if (tokOffset->str() == "(" && (
(tokOffset->link() && Token::Match(tokOffset->link()->previous(), "%type% ) (") &&
Token::Match(tokOffset->link()->next()->link(), ") const|volatile|;")) ||
(Token::simpleMatch(tokOffset, "( (") &&
tokOffset->next() && Token::Match(tokOffset->next()->link()->previous(), "%type% ) (") &&
Token::Match(tokOffset->next()->link()->next()->link(), ") const|volatile| ) ;|,")) ||
(Token::simpleMatch(tokOffset, "( * (") &&
tokOffset->linkAt(2) && Token::Match(tokOffset->linkAt(2)->previous(), "%type% ) (") &&
Token::Match(tokOffset->linkAt(2)->next()->link(), ") const|volatile| ) ;|,")))) {
if (tokOffset->next()->str() == "(")
tokOffset = tokOffset->next();
else if (Token::simpleMatch(tokOffset, "( * (")) {
pointers.push_back("*");
tokOffset = tokOffset->tokAt(2);
}
if (tokOffset->link()->strAt(-2) == "*")
functionPtr = true;
else
function = true;
funcStart = tokOffset->next();
tokOffset = tokOffset->link();
funcEnd = tokOffset->tokAt(-2);
typeName = tokOffset->previous();
argStart = tokOffset->next();
argEnd = tokOffset->next()->link();
if (!argEnd)
syntaxError(argStart);
tok = argEnd->next();
Token *spec = tok;
if (Token::Match(spec, "const|volatile")) {
specStart = spec;
specEnd = spec;
while (Token::Match(spec->next(), "const|volatile")) {
specEnd = spec->next();
spec = specEnd;
}
tok = specEnd->next();
}
if (!tok)
syntaxError(specEnd);
if (tok->str() == ")")
tok = tok->next();
}
else if (Token::Match(tokOffset, "( %type% (")) {
function = true;
if (tokOffset->link()->next()) {
tok = tokOffset->link()->next();
tokOffset = tokOffset->tokAt(2);
typeName = tokOffset->previous();
argStart = tokOffset;
argEnd = tokOffset->link();
} else {
// internal error
continue;
}
}
// pointer to function returning pointer to function
else if (Token::Match(tokOffset, "( * ( * %type% ) (") &&
Token::simpleMatch(tokOffset->linkAt(6), ") ) (") &&
Token::Match(tokOffset->linkAt(6)->linkAt(2), ") ;|,")) {
functionPtrRetFuncPtr = true;
tokOffset = tokOffset->tokAt(6);
typeName = tokOffset->tokAt(-2);
argStart = tokOffset;
argEnd = tokOffset->link();
if (!argEnd)
syntaxError(arrayStart);
argFuncRetStart = argEnd->tokAt(2);
argFuncRetEnd = argFuncRetStart->link();
if (!argFuncRetEnd)
syntaxError(argFuncRetStart);
tok = argFuncRetEnd->next();
}
// function returning pointer to function
else if (Token::Match(tokOffset, "( * %type% (") &&
Token::simpleMatch(tokOffset->linkAt(3), ") ) (") &&
Token::Match(tokOffset->linkAt(3)->linkAt(2), ") ;|,")) {
functionRetFuncPtr = true;
tokOffset = tokOffset->tokAt(3);
typeName = tokOffset->previous();
argStart = tokOffset;
argEnd = tokOffset->link();
argFuncRetStart = argEnd->tokAt(2);
if (!argFuncRetStart)
syntaxError(tokOffset);
argFuncRetEnd = argFuncRetStart->link();
if (!argFuncRetEnd)
syntaxError(tokOffset);
tok = argFuncRetEnd->next();
} else if (Token::Match(tokOffset, "( * ( %type% ) (")) {
functionRetFuncPtr = true;
tokOffset = tokOffset->tokAt(5);
typeName = tokOffset->tokAt(-2);
argStart = tokOffset;
argEnd = tokOffset->link();
if (!argEnd)
syntaxError(arrayStart);
argFuncRetStart = argEnd->tokAt(2);
if (!argFuncRetStart)
syntaxError(tokOffset);
argFuncRetEnd = argFuncRetStart->link();
if (!argFuncRetEnd)
syntaxError(tokOffset);
tok = argFuncRetEnd->next();
}
// pointer/reference to array
else if (Token::Match(tokOffset, "( *|& %type% ) [")) {
ptrToArray = (tokOffset->next()->str() == "*");
refToArray = !ptrToArray;
tokOffset = tokOffset->tokAt(2);
typeName = tokOffset;
arrayStart = tokOffset->tokAt(2);
arrayEnd = arrayStart->link();
if (!arrayEnd)
syntaxError(arrayStart);
tok = arrayEnd->next();
}
// pointer to class member
else if (Token::Match(tokOffset, "( %type% :: * %type% ) ;")) {
tokOffset = tokOffset->tokAt(2);
namespaceStart = tokOffset->previous();
namespaceEnd = tokOffset;
ptrMember = true;
tokOffset = tokOffset->tokAt(2);
typeName = tokOffset;
tok = tokOffset->tokAt(2);
}
// unhandled typedef, skip it and continue
else {
unsupportedTypedef(typeDef);
tok = deleteInvalidTypedef(typeDef);
if (tok == list.front())
//now the next token to process is 'tok', not 'tok->next()';
goback = true;
continue;
}
bool done = false;
bool ok = true;
while (!done) {
std::string pattern = typeName->str();
unsigned int scope = 0;
bool simplifyType = false;
bool inMemberFunc = false;
int memberScope = 0;
bool globalScope = false;
std::size_t classLevel = spaceInfo.size();
for (Token *tok2 = tok; tok2; tok2 = tok2->next()) {
if (mSettings->terminated())
return;
if (tok2->link()) { // Pre-check for performance
// check for end of scope
if (tok2->str() == "}") {
// check for end of member function
if (inMemberFunc) {
--memberScope;
if (memberScope == 0)
inMemberFunc = false;
}
if (classLevel > 0 && tok2 == spaceInfo[classLevel - 1].bodyEnd) {
--classLevel;
pattern.clear();
for (std::size_t i = classLevel; i < spaceInfo.size(); ++i)
pattern += (spaceInfo[i].className + " :: ");
pattern += typeName->str();
} else {
if (scope == 0)
break;
--scope;
}
}
// check for member functions
else if (isCPP() && Token::Match(tok2, ")|] const| {")) {
const Token *temp = tok2;
while (temp && temp->str() == "]" && temp->link() && temp->link()->previous())
temp = temp->link()->previous();
if (!temp || !temp->link() || !temp->link()->previous())
continue;
const Token *func = temp->link()->previous();
if (temp->str() != ")")
continue;
if (!func || !func->previous()) // Ticket #4239
continue;
/** @todo add support for multi-token operators */
if (func->previous()->str() == "operator")
func = func->previous();
if (!func->previous())
syntaxError(func);
// check for qualifier
if (func->previous()->str() == "::") {
// check for available and matching class name
if (!spaceInfo.empty() && classLevel < spaceInfo.size() &&
func->strAt(-2) == spaceInfo[classLevel].className) {
memberScope = 0;
inMemberFunc = true;
}
}
}
// check for entering a new scope
else if (tok2->str() == "{") {
// check for entering a new namespace
if (isCPP() && tok2->strAt(-2) == "namespace") {
if (classLevel < spaceInfo.size() &&
spaceInfo[classLevel].isNamespace &&
spaceInfo[classLevel].className == tok2->previous()->str()) {
spaceInfo[classLevel].bodyEnd = tok2->link();
++classLevel;
pattern.clear();
for (std::size_t i = classLevel; i < spaceInfo.size(); ++i)
pattern += (spaceInfo[i].className + " :: ");
pattern += typeName->str();
}
++scope;
}
// keep track of scopes within member function
if (inMemberFunc)
++memberScope;
++scope;
}
}
// check for operator typedef
/** @todo add support for multi-token operators */
else if (isCPP() &&
tok2->str() == "operator" &&
tok2->next() &&
tok2->next()->str() == typeName->str() &&
tok2->linkAt(2) &&
tok2->strAt(2) == "(" &&
Token::Match(tok2->linkAt(2), ") const| {")) {
// check for qualifier
if (tok2->previous()->str() == "::") {
// check for available and matching class name
if (!spaceInfo.empty() && classLevel < spaceInfo.size() &&
tok2->strAt(-2) == spaceInfo[classLevel].className) {
tok2 = tok2->next();
simplifyType = true;
}
}
}
// check for typedef that can be substituted
else if (Token::simpleMatch(tok2, pattern.c_str()) ||
(inMemberFunc && tok2->str() == typeName->str())) {
// member function class variables don't need qualification
if (!(inMemberFunc && tok2->str() == typeName->str()) && pattern.find("::") != std::string::npos) { // has a "something ::"
Token *start = tok2;
std::size_t count = 0;
int back = int(classLevel) - 1;
bool good = true;
// check for extra qualification
while (back >= 0) {
Token *qualificationTok = start->tokAt(-2);
if (!Token::Match(qualificationTok, "%type% ::"))
break;
if (qualificationTok->str() == spaceInfo[back].className) {
start = qualificationTok;
back--;
count++;
} else {
good = false;
break;
}
}
// check global namespace
if (good && back == 0 && start->strAt(-1) == "::")
good = false;
if (good) {
// remove any extra qualification if present
while (count) {
tok2->tokAt(-3)->deleteNext(2);
--count;
}
// remove global namespace if present
if (tok2->strAt(-1) == "::") {
tok2->tokAt(-2)->deleteNext();
globalScope = true;
}
// remove qualification if present
for (std::size_t i = classLevel; i < spaceInfo.size(); ++i) {
tok2->deleteNext(2);
}
simplifyType = true;
}
} else {
if (tok2->strAt(-1) == "::") {
std::size_t relativeSpaceInfoSize = spaceInfo.size();
Token * tokBeforeType = tok2->previous();
while (relativeSpaceInfoSize != 0 &&
tokBeforeType && tokBeforeType->str() == "::" &&
tokBeforeType->strAt(-1) == spaceInfo[relativeSpaceInfoSize-1].className) {
tokBeforeType = tokBeforeType->tokAt(-2);
--relativeSpaceInfoSize;
}
if (tokBeforeType && tokBeforeType->str() != "::") {
Token::eraseTokens(tokBeforeType, tok2);
simplifyType = true;
}
} else if (Token::Match(tok2->previous(), "case|;|{|} %type% :")) {
tok2 = tok2->next();
} else if (duplicateTypedef(&tok2, typeName, typeDef)) {
// skip to end of scope if not already there
if (tok2->str() != "}") {
while (tok2->next()) {
if (tok2->next()->str() == "{")
tok2 = tok2->linkAt(1)->previous();
else if (tok2->next()->str() == "}")
break;
tok2 = tok2->next();
}
}
} else if (Token::Match(tok2->tokAt(-2), "%type% *|&")) {
// Ticket #5868: Don't substitute variable names
} else if (tok2->previous()->str() != ".") {
simplifyType = true;
}
}
}
if (simplifyType) {
// can't simplify 'operator functionPtr ()' and 'functionPtr operator ... ()'
if (functionPtr && (tok2->previous()->str() == "operator" ||
(tok2->next() && tok2->next()->str() == "operator"))) {
simplifyType = false;
tok2 = tok2->next();
continue;
}
// There are 2 categories of typedef substitutions:
// 1. variable declarations that preserve the variable name like
// global, local, and function parameters
// 2. not variable declarations that have no name like derived
// classes, casts, operators, and template parameters
// try to determine which category this substitution is
bool inCast = false;
bool inTemplate = false;
bool inOperator = false;
bool inSizeof = false;
const bool sameStartEnd = (typeStart == typeEnd);
// check for derived class: class A : some_typedef {
const bool isDerived = Token::Match(tok2->previous(), "public|protected|private %type% {|,");
// check for cast: (some_typedef) A or static_cast<some_typedef>(A)
// todo: check for more complicated casts like: (const some_typedef *)A
if ((tok2->previous()->str() == "(" && tok2->next()->str() == ")" && tok2->strAt(-2) != "sizeof") ||
(tok2->previous()->str() == "<" && Token::simpleMatch(tok2->next(), "> (")) ||
Token::Match(tok2->tokAt(-2), "( const %name% )"))
inCast = true;
// check for template parameters: t<some_typedef> t1
else if (Token::Match(tok2->previous(), "<|,") &&
Token::Match(tok2->next(), "&|*| &|*| >|,"))
inTemplate = true;
else if (Token::Match(tok2->tokAt(-2), "sizeof ( %type% )"))
inSizeof = true;
// check for operator
if (tok2->strAt(-1) == "operator" ||
Token::simpleMatch(tok2->tokAt(-2), "operator const"))
inOperator = true;
if (typeStart->str() == "typename" && tok2->strAt(-1)=="typename") {
// Remove one typename if it is already contained in the goal
typeStart = typeStart->next();
}
// skip over class or struct in derived class declaration
bool structRemoved = false;
if (isDerived && Token::Match(typeStart, "class|struct")) {
if (typeStart->str() == "struct")
structRemoved = true;
typeStart = typeStart->next();
}
if (Token::Match(typeStart, "struct|class") && Token::Match(tok2, "%name% ::"))
typeStart = typeStart->next();
if (sameStartEnd)
typeEnd = typeStart;
// start substituting at the typedef name by replacing it with the type
tok2->str(typeStart->str());
// restore qualification if it was removed
if (typeStart->str() == "struct" || structRemoved) {
if (structRemoved)
tok2 = tok2->previous();
if (globalScope) {
tok2->insertToken("::");
tok2 = tok2->next();
}
for (std::size_t i = classLevel; i < spaceInfo.size(); ++i) {
tok2->insertToken(spaceInfo[i].className);
tok2 = tok2->next();
tok2->insertToken("::");
tok2 = tok2->next();
}
}
// add remainder of type
tok2 = TokenList::copyTokens(tok2, typeStart->next(), typeEnd);
if (!pointers.empty()) {
for (const std::string &p : pointers) {
tok2->insertToken(p);
tok2 = tok2->next();
}
}
if (funcStart && funcEnd) {
tok2->insertToken("(");
tok2 = tok2->next();
Token *tok3 = tok2;
tok2 = TokenList::copyTokens(tok2, funcStart, funcEnd);
if (!inCast)
tok2 = processFunc(tok2, inOperator);
if (!tok2)
break;
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, tok3);
tok2 = TokenList::copyTokens(tok2, argStart, argEnd);
if (specStart) {
Token *spec = specStart;
tok2->insertToken(spec->str());
tok2 = tok2->next();
while (spec != specEnd) {
spec = spec->next();
tok2->insertToken(spec->str());
tok2 = tok2->next();
}
}
}
else if (functionPtr || function) {
// don't add parentheses around function names because it
// confuses other simplifications
bool needParen = true;
if (!inTemplate && function && tok2->next() && tok2->next()->str() != "*")
needParen = false;
if (needParen) {
tok2->insertToken("(");
tok2 = tok2->next();
}
Token *tok3 = tok2;
if (namespaceStart) {
const Token *tok4 = namespaceStart;
while (tok4 != namespaceEnd) {
tok2->insertToken(tok4->str());
tok2 = tok2->next();
tok4 = tok4->next();
}
tok2->insertToken(namespaceEnd->str());
tok2 = tok2->next();
}
if (functionPtr) {
tok2->insertToken("*");
tok2 = tok2->next();
}
if (!inCast)
tok2 = processFunc(tok2, inOperator);
if (needParen) {
if (!tok2)
syntaxError(nullptr);
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, tok3);
}
if (!tok2)
syntaxError(nullptr);
tok2 = TokenList::copyTokens(tok2, argStart, argEnd);
if (inTemplate) {
if (!tok2)
syntaxError(nullptr);
tok2 = tok2->next();
}
if (specStart) {
Token *spec = specStart;
tok2->insertToken(spec->str());
tok2 = tok2->next();
while (spec != specEnd) {
spec = spec->next();
tok2->insertToken(spec->str());
tok2 = tok2->next();
}
}
} else if (functionRetFuncPtr || functionPtrRetFuncPtr) {
tok2->insertToken("(");
tok2 = tok2->next();
Token *tok3 = tok2;
tok2->insertToken("*");
tok2 = tok2->next();
Token * tok4 = nullptr;
if (functionPtrRetFuncPtr) {
tok2->insertToken("(");
tok2 = tok2->next();
tok4 = tok2;
tok2->insertToken("*");
tok2 = tok2->next();
}
// skip over variable name if there
if (!inCast) {
if (!tok2 || !tok2->next())
syntaxError(nullptr);
if (tok2->next()->str() != ")")
tok2 = tok2->next();
}
if (tok4 && functionPtrRetFuncPtr) {
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, tok4);
}
tok2 = TokenList::copyTokens(tok2, argStart, argEnd);
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, tok3);
tok2 = TokenList::copyTokens(tok2, argFuncRetStart, argFuncRetEnd);
} else if (ptrToArray || refToArray) {
tok2->insertToken("(");
tok2 = tok2->next();
Token *tok3 = tok2;
if (ptrToArray)
tok2->insertToken("*");
else
tok2->insertToken("&");
tok2 = tok2->next();
// skip over name
if (tok2->next() && tok2->next()->str() != ")") {
if (tok2->next()->str() != "(")
tok2 = tok2->next();
// check for function and skip over args
if (tok2 && tok2->next() && tok2->next()->str() == "(")
tok2 = tok2->next()->link();
// check for array
if (tok2 && tok2->next() && tok2->next()->str() == "[")
tok2 = tok2->next()->link();
} else {
// syntax error
}
tok2->insertToken(")");
Token::createMutualLinks(tok2->next(), tok3);
} else if (ptrMember) {
if (Token::simpleMatch(tok2, "* (")) {
tok2->insertToken("*");
tok2 = tok2->next();
} else {
// This is the case of casting operator.
// Name is not available, and () should not be
// inserted
const bool castOperator = inOperator && Token::Match(tok2, "%type% (");
Token *openParenthesis = nullptr;
if (!castOperator) {
tok2->insertToken("(");
tok2 = tok2->next();
openParenthesis = tok2;
}
const Token *tok4 = namespaceStart;
while (tok4 != namespaceEnd) {
tok2->insertToken(tok4->str());
tok2 = tok2->next();
tok4 = tok4->next();
}
tok2->insertToken(namespaceEnd->str());
tok2 = tok2->next();
tok2->insertToken("*");
tok2 = tok2->next();
if (openParenthesis) {
// Skip over name, if any
if (Token::Match(tok2->next(), "%name%"))
tok2 = tok2->next();
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, openParenthesis);
}
}
} else if (typeOf) {
tok2 = TokenList::copyTokens(tok2, argStart, argEnd);
} else if (Token::Match(tok2, "%name% [")) {
while (Token::Match(tok2, "%name%|] [")) {
tok2 = tok2->linkAt(1);
}
tok2 = tok2->previous();
}
if (arrayStart && arrayEnd) {
do {
if (!tok2->next())
syntaxError(tok2); // can't recover so quit
if (!inCast && !inSizeof && !inTemplate)
tok2 = tok2->next();
if (tok2->str() == "const")
tok2 = tok2->next();
// reference or pointer to array?
if (tok2->str() == "&" || tok2->str() == "*") {
tok2 = tok2->previous();
tok2->insertToken("(");
Token *tok3 = tok2->next();
// handle missing variable name
if (tok2->strAt(3) == ")" || tok2->strAt(3) == ",")
tok2 = tok2->tokAt(2);
else
tok2 = tok2->tokAt(3);
if (!tok2)
syntaxError(nullptr);
while (tok2->strAt(1) == "::")
tok2 = tok2->tokAt(2);
// skip over function parameters
if (tok2->strAt(1) == "(") {
tok2 = tok2->linkAt(1);
if (tok2->strAt(1) == "const")
tok2 = tok2->next();
}
tok2->insertToken(")");
tok2 = tok2->next();
Token::createMutualLinks(tok2, tok3);
}
if (!tok2->next())
syntaxError(tok2); // can't recover so quit
// skip over array dimensions
while (tok2->next()->str() == "[")
tok2 = tok2->linkAt(1);
tok2 = TokenList::copyTokens(tok2, arrayStart, arrayEnd);
if (!tok2->next())
syntaxError(tok2);
if (tok2->str() == "=") {
if (!tok2->next())
syntaxError(tok2);
if (tok2->next()->str() == "{")
tok2 = tok2->next()->link()->next();
else if (tok2->next()->str().at(0) == '\"')
tok2 = tok2->tokAt(2);
}
} while (Token::Match(tok2, ", %name% ;|=|,"));
}
simplifyType = false;
}
if (!tok2)
break;
}
if (!tok)
syntaxError(nullptr);
if (tok->str() == ";")
done = true;
else if (tok->str() == ",") {
arrayStart = nullptr;
arrayEnd = nullptr;
tokOffset = tok->next();
pointers.clear();
while (Token::Match(tokOffset, "*|&")) {
pointers.push_back(tokOffset->str());
tokOffset = tokOffset->next();
}
if (Token::Match(tokOffset, "%type%")) {
typeName = tokOffset;
tokOffset = tokOffset->next();
if (tokOffset && tokOffset->str() == "[") {
arrayStart = tokOffset;
for (;;) {
while (tokOffset->next() && !Token::Match(tokOffset->next(), ";|,"))
tokOffset = tokOffset->next();
if (!tokOffset->next())
return; // invalid input
else if (tokOffset->next()->str() == ";")
break;
else if (tokOffset->str() == "]")
break;
else
tokOffset = tokOffset->next();
}
arrayEnd = tokOffset;
tokOffset = tokOffset->next();
}
if (Token::Match(tokOffset, ";|,"))
tok = tokOffset;
else {
// we encountered a typedef we don't support yet so just continue
done = true;
ok = false;
}
} else {
// we encountered a typedef we don't support yet so just continue
done = true;
ok = false;
}
} else {
// something is really wrong (internal error)
done = true;
ok = false;
}
}
if (ok) {
// remove typedef
Token::eraseTokens(typeDef, tok);
if (typeDef != list.front()) {
tok = typeDef->previous();
tok->deleteNext();
//no need to remove last token in the list
if (tok->tokAt(2))
tok->deleteNext();
} else {
list.front()->deleteThis();
//no need to remove last token in the list
if (list.front()->next())
list.front()->deleteThis();
tok = list.front();
//now the next token to process is 'tok', not 'tok->next()';
goback = true;
}
}
}
}
namespace {
struct ScopeInfo3 {
ScopeInfo3(const std::string &name_, const Token *bodyEnd_) : name(name_), bodyEnd(bodyEnd_) {}
const std::string name;
const Token * const bodyEnd;
std::set<std::string> usingNamespaces;
};
std::string getScopeName(const std::list<ScopeInfo3> &scopeInfo)
{
std::string ret;
for (const ScopeInfo3 &i : scopeInfo) {
if (!i.name.empty())
ret += (ret.empty() ? "" : " :: ") + i.name;
}
return ret;
}
void setScopeInfo(Token *tok, std::list<ScopeInfo3> *scopeInfo, bool all = false)
{
while (tok->str() == "}" && !scopeInfo->empty() && tok == scopeInfo->back().bodyEnd)
scopeInfo->pop_back();
if (!Token::Match(tok, "namespace|class|struct|union %name% {|:|::")) {
// check for using namespace
if (Token::Match(tok, "using namespace %name% ;|::")) {
const Token * tok1 = tok->tokAt(2);
std::string nameSpace;
while (tok1 && tok1->str() != ";") {
if (!nameSpace.empty())
nameSpace += " ";
nameSpace += tok1->str();
tok1 = tok1->next();
}
scopeInfo->back().usingNamespaces.insert(nameSpace);
}
// check for member function
else if (tok->str() == "{") {
bool added = false;
Token *tok1 = tok;
while (Token::Match(tok1->previous(), "const|volatile|final|override|&|&&|noexcept"))
tok1 = tok1->previous();
if (tok1 && tok1->previous() && tok1->strAt(-1) == ")") {
tok1 = tok1->linkAt(-1);
if (Token::Match(tok1->previous(), "throw|noexcept")) {
tok1 = tok1->previous();
while (Token::Match(tok1->previous(), "const|volatile|final|override|&|&&|noexcept"))
tok1 = tok1->previous();
if (tok1->strAt(-1) != ")")
return;
} else if (Token::Match(tok->tokAt(-2), ":|, %name%")) {
tok1 = tok1->tokAt(-2);
if (tok1->strAt(-1) != ")")
return;
}
if (tok1->strAt(-1) == ">")
tok1 = tok1->previous()->findOpeningBracket();
if (tok1 && Token::Match(tok1->tokAt(-3), "%name% :: %name%")) {
tok1 = tok1->tokAt(-2);
std::string scope = tok1->strAt(-1);
while (Token::Match(tok1->tokAt(-2), ":: %name%")) {
scope = tok1->strAt(-3) + " :: " + scope;
tok1 = tok1->tokAt(-2);
}
scopeInfo->emplace_back(scope, tok->link());
added = true;
}
}
if (all && !added)
scopeInfo->emplace_back("", tok->link());
}
return;
}
tok = tok->next();
std::string classname = tok->str();
while (Token::Match(tok, "%name% :: %name%")) {
tok = tok->tokAt(2);
classname += " :: " + tok->str();
}
tok = tok->next();
if (tok && tok->str() == ":") {
while (tok && !Token::Match(tok, ";|{"))
tok = tok->next();
}
if (tok && tok->str() == "{") {
scopeInfo->emplace_back(classname,tok->link());
}
}
Token *findSemicolon(Token *tok)
{
unsigned int level = 0;
for (; tok && (level > 0 || tok->str() != ";"); tok = tok->next()) {
if (tok->str() == "{")
++level;
else if (level > 0 && tok->str() == "}")
--level;
}
return tok;
}
bool usingMatch(
const Token *nameToken,
const std::string &scope,
Token **tok,
const std::string &scope1,
const std::list<ScopeInfo3> &scopeList1)
{
Token *tok1 = *tok;
if (tok1 && tok1->str() != nameToken->str())
return false;
// skip this using
if (tok1 == nameToken) {
*tok = findSemicolon(tok1);
return false;
}
// skip other using with this name
if (tok1->strAt(-1) == "using") {
// fixme: this is wrong
// skip to end of scope
if (scopeList1.back().bodyEnd)
*tok = scopeList1.back().bodyEnd->previous();
return false;
}
if (Token::Match(tok1->tokAt(-1), "struct|union|enum")) {
// fixme
return false;
}
// get qualification
std::string qualification;
const Token* tok2 = tok1;
std::string::size_type index = scope.size();
std::string::size_type new_index = std::string::npos;
bool match = true;
while (tok2->strAt(-1) == "::") {
std::string last;
if (match && !scope1.empty()) {
new_index = scope1.rfind(' ', index - 1);
if (new_index != std::string::npos)
last = scope1.substr(new_index, index - new_index);
else if (!qualification.empty())
last.clear();
else
last = scope1;
} else
match = false;
if (match && tok2->strAt(-2) == last)
index = new_index;
else {
if (!qualification.empty())
qualification = " :: " + qualification;
qualification = tok2->strAt(-2) + qualification;
}
tok2 = tok2->tokAt(-2);
}
// todo: check using namespace
std::string fullScope1 = scope1;
if (!scope1.empty() && !qualification.empty())
fullScope1 += " :: ";
fullScope1 += qualification;
if (scope == fullScope1)
return true;
std::string newScope1 = scope1;
// scopes didn't match so try higher scopes
while (!newScope1.empty()) {
std::string::size_type separator = newScope1.rfind(" :: ", index - 1);
if (separator != std::string::npos)
newScope1 = newScope1.substr(0, separator);
else
newScope1.clear();
std::string newFullScope1 = newScope1;
if (!newScope1.empty() && !qualification.empty())
newFullScope1 += " :: ";
newFullScope1 += qualification;
if (scope == newFullScope1)
return true;
}
return false;
}
} // namespace
bool Tokenizer::simplifyUsing()
{
bool substitute = false;
std::list<ScopeInfo3> scopeList;
bool inTemplateDefinition = false;
const Token *endOfTemplateDefinition = nullptr;
bool isVariable = false;
struct Using {
Using(Token *start, Token *end) : startTok(start), endTok(end) { }
Token *startTok;
Token *endTok;
};
std::list<Using> usingList;
scopeList.emplace_back("", nullptr);
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (mErrorLogger && !list.getFiles().empty())
mErrorLogger->reportProgress(list.getFiles()[0], "Tokenize (using)", tok->progressValue());
if (mSettings->terminated())
return substitute;
if (Token::Match(tok, "{|}|namespace|class|struct|union") ||
Token::Match(tok, "using namespace %name% ;|::")) {
setScopeInfo(tok, &scopeList);
}
if (inTemplateDefinition) {
if (!endOfTemplateDefinition) {
if (isVariable)
endOfTemplateDefinition = findSemicolon(tok);
else if (tok->str() == "{")
endOfTemplateDefinition = tok->link();
}
if (tok == endOfTemplateDefinition) {
inTemplateDefinition = false;
endOfTemplateDefinition = nullptr;
continue;
}
}
if (tok->str()=="template") {
if (Token::Match(tok->next(), "< !!>"))
inTemplateDefinition = true;
else
inTemplateDefinition = false;
}
if (!inTemplateDefinition) {
// look for non-template type aliases
if (tok->strAt(-1) != ">" &&
(Token::Match(tok, "using %name% = ::| %name%") ||
(Token::Match(tok, "using %name% [ [") &&
Token::Match(tok->linkAt(2), "] ] = ::| %name%")))) {
std::list<ScopeInfo3> scopeList1;
scopeList1.emplace_back("", nullptr);
std::string name = tok->strAt(1);
const Token *nameToken = tok->next();
std::string scope = getScopeName(scopeList);
Token *usingStart = tok;
Token *start;
if (tok->strAt(2) == "=")
start = tok->tokAt(3);
else
start = tok->linkAt(2)->tokAt(3);
Token *usingEnd = findSemicolon(start);
if (!usingEnd)
continue;
// Move struct defined in using out of using.
// using T = struct t { }; => struct t { }; using T = struct t;
// fixme: this doesn't handle attributes
if (Token::Match(start, "struct|union|enum %name%| {")) {
if (start->strAt(1) != "{") {
Token *structEnd = start->linkAt(2);
structEnd->insertToken(";", "");
list.copyTokens(structEnd->next(), tok, start->next());
usingStart = structEnd->tokAt(2);
nameToken = usingStart->next();
if (usingStart->strAt(2) == "=")
start = usingStart->tokAt(3);
else
start = usingStart->linkAt(2)->tokAt(3);
usingEnd = findSemicolon(start);
tok->deleteThis();
tok->deleteThis();
tok->deleteThis();
tok = usingStart;
} else {
Token *structEnd = start->linkAt(1);
structEnd->insertToken(";", "");
std::string newName;
if (structEnd->strAt(2) == ";")
newName = name;
else
newName = "Unnamed" + MathLib::toString(mUnnamedCount++);
list.copyTokens(structEnd->next(), tok, start);
structEnd->tokAt(5)->insertToken(newName, "");
start->insertToken(newName, "");
usingStart = structEnd->tokAt(2);
nameToken = usingStart->next();
if (usingStart->strAt(2) == "=")
start = usingStart->tokAt(3);
else
start = usingStart->linkAt(2)->tokAt(3);
usingEnd = findSemicolon(start);
tok->deleteThis();
tok->deleteThis();
tok->deleteThis();
tok = usingStart;
}
}
// Unfortunately we have to start searching from the beginning
// of the token stream because templates are instantiated at
// the end of the token stream and it may be used before then.
std::string scope1;
bool skip = false; // don't erase type aliases we can't parse
for (Token* tok1 = list.front(); tok1; tok1 = tok1->next()) {
if ((Token::Match(tok1, "{|}|namespace|class|struct|union") && tok1->strAt(-1) != "using") ||
Token::Match(tok1, "using namespace %name% ;|::")) {
setScopeInfo(tok1, &scopeList1, true);
scope1 = getScopeName(scopeList1);
continue;
}
if (!usingMatch(nameToken, scope, &tok1, scope1, scopeList1))
continue;
// remove the qualification
while (tok1->strAt(-1) == "::" && tok1->strAt(-2) == scope) {
tok1->deletePrevious();
tok1->deletePrevious();
}
Token * arrayStart = nullptr;
// parse the type
Token *type = start;
if (type->str() == "::") {
type = type->next();
while (Token::Match(type, "%type% ::"))
type = type->tokAt(2);
if (Token::Match(type, "%type%"))
type = type->next();
} else if (Token::Match(type, "%type% ::")) {
do {
type = type->tokAt(2);
} while (Token::Match(type, "%type% ::"));
if (Token::Match(type, "%type%"))
type = type->next();
} else if (Token::Match(type, "%type%")) {
while (Token::Match(type, "const|struct|union|enum %type%") ||
(type->next() && type->next()->isStandardType()))
type = type->next();
type = type->next();
while (Token::Match(type, "%type%") &&
(type->isStandardType() || Token::Match(type, "unsigned|signed"))) {
type = type->next();
}
bool atEnd = false;
while (!atEnd) {
if (type && type->str() == "::") {
type = type->next();
}
if (Token::Match(type, "%type%") &&
type->next() && !Token::Match(type->next(), "[|;|,|(")) {
type = type->next();
} else if (Token::simpleMatch(type, "const (")) {
type = type->next();
atEnd = true;
} else
atEnd = true;
}
} else
syntaxError(type);
// check for invalid input
if (!type)
syntaxError(tok1);
// check for template
if (type->str() == "<") {
type = type->findClosingBracket();
while (type && Token::Match(type->next(), ":: %type%"))
type = type->tokAt(2);
if (!type) {
syntaxError(tok1);
}
while (Token::Match(type->next(), "const|volatile"))
type = type->next();
type = type->next();
}
// check for pointers and references
std::list<std::string> pointers;
while (Token::Match(type, "*|&|&&|const")) {
pointers.push_back(type->str());
type = type->next();
}
// check for array
if (type && type->str() == "[") {
do {
if (!arrayStart)
arrayStart = type;
bool atEnd = false;
while (!atEnd) {
while (type->next() && !Token::Match(type->next(), ";|,")) {
type = type->next();
}
if (!type->next())
syntaxError(type); // invalid input
else if (type->next()->str() == ";")
atEnd = true;
else if (type->str() == "]")
atEnd = true;
else
type = type->next();
}
type = type->next();
} while (type && type->str() == "[");
}
Token* after = tok1->next();
// check if type was parsed
if (type && type == usingEnd) {
// check for array syntax and add type around variable
if (arrayStart) {
if (Token::Match(tok1->next(), "%name%")) {
list.copyTokens(tok1->next(), arrayStart, usingEnd->previous());
list.copyTokens(tok1, start, arrayStart->previous());
tok1->deleteThis();
substitute = true;
}
} else {
// just replace simple type aliases
list.copyTokens(tok1, start, usingEnd->previous());
tok1->deleteThis();
substitute = true;
}
} else {
skip = true;
if (mSettings->debugwarnings && mErrorLogger) {
std::string str;
for (Token *tok3 = usingStart; tok3 && tok3 != usingEnd; tok3 = tok3->next()) {
if (!str.empty())
str += ' ';
str += tok3->str();
}
str += " ;";
std::list<const Token *> callstack(1, usingStart);
mErrorLogger->reportErr(ErrorLogger::ErrorMessage(callstack, &list, Severity::debug, "debug",
"Failed to parse \'" + str + "\'. The checking continues anyway.", false));
}
}
tok1 = after;
}
if (!skip)
usingList.emplace_back(usingStart, usingEnd);
}
}
}
// delete all used type alias definitions
for (std::list<Using>::reverse_iterator it = usingList.rbegin(); it != usingList.rend(); ++it) {
Token *usingStart = it->startTok;
Token *usingEnd = it->endTok;
if (usingStart->previous()) {
if (usingEnd->next())
Token::eraseTokens(usingStart->previous(), usingEnd->next());
else {
Token::eraseTokens(usingStart->previous(), usingEnd);
usingEnd->deleteThis();
}
} else {
if (usingEnd->next()) {
Token::eraseTokens(usingStart, usingEnd->next());
usingStart->deleteThis();
} else {
// this is the only code being checked so leave ';'
Token::eraseTokens(usingStart, usingEnd);
usingStart->deleteThis();
}
}
}
return substitute;
}
void Tokenizer::simplifyMulAndParens()
{
if (!list.front())
return;
for (Token *tok = list.front()->tokAt(3); tok; tok = tok->next()) {
if (!tok->isName())
continue;
//fix ticket #2784 - improved by ticket #3184
unsigned int closedPars = 0;
Token *tokend = tok->next();
Token *tokbegin = tok->previous();
while (tokend && tokend->str() == ")") {
++closedPars;
tokend = tokend->next();
}
if (!tokend || !(tokend->isAssignmentOp()))
continue;
while (Token::Match(tokbegin, "&|(")) {
if (tokbegin->str() == "&") {
if (Token::Match(tokbegin->tokAt(-2), "[;{}&(] *")) {
//remove '* &'
tokbegin = tokbegin->tokAt(-2);
tokbegin->deleteNext(2);
} else if (Token::Match(tokbegin->tokAt(-3), "[;{}&(] * (")) {
if (closedPars == 0)
break;
--closedPars;
//remove ')'
tok->deleteNext();
//remove '* ( &'
tokbegin = tokbegin->tokAt(-3);
tokbegin->deleteNext(3);
} else
break;
} else if (tokbegin->str() == "(") {
if (closedPars == 0)
break;
//find consecutive opening parentheses
unsigned int openPars = 0;
while (tokbegin && tokbegin->str() == "(" && openPars <= closedPars) {
++openPars;
tokbegin = tokbegin->previous();
}
if (!tokbegin || openPars > closedPars)
break;
if ((openPars == closedPars && Token::Match(tokbegin, "[;{}]")) ||
Token::Match(tokbegin->tokAt(-2), "[;{}&(] * &") ||
Token::Match(tokbegin->tokAt(-3), "[;{}&(] * ( &")) {
//remove the excessive parentheses around the variable
while (openPars > 0) {
tok->deleteNext();
tokbegin->deleteNext();
--closedPars;
--openPars;
}
} else
break;
}
}
}
}
bool Tokenizer::createTokens(std::istream &code,
const std::string& FileName)
{
// make sure settings specified
assert(mSettings);
return list.createTokens(code, FileName);
}
void Tokenizer::createTokens(const simplecpp::TokenList *tokenList)
{
// make sure settings specified
assert(mSettings);
list.createTokens(tokenList);
}
bool Tokenizer::simplifyTokens1(const std::string &configuration)
{
// Fill the map mTypeSize..
fillTypeSizes();
mConfiguration = configuration;
if (!simplifyTokenList1(list.getFiles().front().c_str()))
return false;
if (mTimerResults) {
Timer t("Tokenizer::simplifyTokens1::createAst", mSettings->showtime, mTimerResults);
list.createAst();
list.validateAst();
} else {
list.createAst();
list.validateAst();
}
if (mTimerResults) {
Timer t("Tokenizer::simplifyTokens1::createSymbolDatabase", mSettings->showtime, mTimerResults);
createSymbolDatabase();
} else {
createSymbolDatabase();
}
// Use symbol database to identify rvalue references. Split && to & &. This is safe, since it doesn't delete any tokens (which might be referenced by symbol database)
for (const Variable* var : mSymbolDatabase->variableList()) {
if (var && var->isRValueReference()) {
Token* endTok = const_cast<Token*>(var->typeEndToken());
endTok->str("&");
endTok->astOperand1(nullptr);
endTok->astOperand2(nullptr);
endTok->insertToken("&");
endTok->next()->scope(endTok->scope());
}
}
if (mTimerResults) {
Timer t("Tokenizer::simplifyTokens1::setValueType", mSettings->showtime, mTimerResults);
mSymbolDatabase->setValueTypeInTokenList();
} else {
mSymbolDatabase->setValueTypeInTokenList();
}
if (mTimerResults) {
Timer t("Tokenizer::simplifyTokens1::ValueFlow", mSettings->showtime, mTimerResults);
ValueFlow::setValues(&list, mSymbolDatabase, mErrorLogger, mSettings);
} else {
ValueFlow::setValues(&list, mSymbolDatabase, mErrorLogger, mSettings);
}
mSymbolDatabase->setArrayDimensionsUsingValueFlow();
printDebugOutput(1);
return true;
}
bool Tokenizer::tokenize(std::istream &code,
const char FileName[],
const std::string &configuration)
{
if (!createTokens(code, FileName))
return false;
return simplifyTokens1(configuration);
}
//---------------------------------------------------------------------------
void Tokenizer::findComplicatedSyntaxErrorsInTemplates()
{
validate();
mTemplateSimplifier->checkComplicatedSyntaxErrorsInTemplates();
}
void Tokenizer::checkForEnumsWithTypedef()
{
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "enum %name% {")) {
tok = tok->tokAt(2);
const Token *tok2 = Token::findsimplematch(tok, "typedef", tok->link());
if (tok2)
syntaxError(tok2);
tok = tok->link();
}
}
}
void Tokenizer::fillTypeSizes()
{
mTypeSize.clear();
mTypeSize["char"] = 1;
mTypeSize["_Bool"] = mSettings->sizeof_bool;
mTypeSize["bool"] = mSettings->sizeof_bool;
mTypeSize["short"] = mSettings->sizeof_short;
mTypeSize["int"] = mSettings->sizeof_int;
mTypeSize["long"] = mSettings->sizeof_long;
mTypeSize["float"] = mSettings->sizeof_float;
mTypeSize["double"] = mSettings->sizeof_double;
mTypeSize["wchar_t"] = mSettings->sizeof_wchar_t;
mTypeSize["size_t"] = mSettings->sizeof_size_t;
mTypeSize["*"] = mSettings->sizeof_pointer;
}
void Tokenizer::combineOperators()
{
const bool cpp = isCPP();
// Combine tokens..
for (Token *tok = list.front(); tok && tok->next(); tok = tok->next()) {
const char c1 = tok->str()[0];
if (tok->str().length() == 1 && tok->next()->str().length() == 1) {
const char c2 = tok->next()->str()[0];
// combine +-*/ and =
if (c2 == '=' && (std::strchr("+-*/%|^=!<>", c1))) {
// skip templates
if (cpp && tok->str() == ">") {
const Token *opening = tok->findOpeningBracket();
if (opening) {
if (Token::Match(opening->previous(), "%name%"))
continue;
}
}
tok->str(tok->str() + c2);
tok->deleteNext();
continue;
}
} else if (tok->next()->str() == "=") {
if (tok->str() == ">>") {
tok->str(">>=");
tok->deleteNext();
} else if (tok->str() == "<<") {
tok->str("<<=");
tok->deleteNext();
}
} else if (cpp && (c1 == 'p' || c1 == '_') &&
Token::Match(tok, "private|protected|public|__published : !!:")) {
bool simplify = false;
unsigned int par = 0U;
for (const Token *prev = tok->previous(); prev; prev = prev->previous()) {
if (prev->str() == ")") {
++par;
} else if (prev->str() == "(") {
if (par == 0U)
break;
--par;
}
if (par != 0U || prev->str() == "(")
continue;
if (Token::Match(prev, "[;{}]")) {
simplify = true;
break;
}
if (prev->isName() && prev->isUpperCaseName())
continue;
if (prev->isName() && endsWith(prev->str(), ':'))
simplify = true;
break;
}
if (simplify) {
tok->str(tok->str() + ":");
tok->deleteNext();
}
} else if (tok->str() == "->") {
// If the preceding sequence is "( & %name% )", replace it by "%name%"
Token *t = tok->tokAt(-4);
if (Token::Match(t, "( & %name% )")) {
t->deleteThis();
t->deleteThis();
t->deleteNext();
}
tok->str(".");
tok->originalName("->");
}
}
}
void Tokenizer::combineStringAndCharLiterals()
{
// Combine strings
for (Token *tok = list.front();
tok;
tok = tok->next()) {
if (tok->str()[0] != '"')
continue;
tok->str(simplifyString(tok->str()));
while (tok->next() && tok->next()->tokType() == Token::eString) {
// Two strings after each other, combine them
tok->concatStr(simplifyString(tok->next()->str()));
tok->deleteNext();
}
}
}
void Tokenizer::concatenateNegativeNumberAndAnyPositive()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "?|:|,|(|[|{|return|case|sizeof|%op% +|-") || tok->tokType() == Token::eIncDecOp)
continue;
while (tok->str() != ">" && tok->next() && tok->next()->str() == "+")
tok->deleteNext();
if (Token::Match(tok->next(), "- %num%")) {
tok->deleteNext();
tok->next()->str("-" + tok->next()->str());
}
}
}
void Tokenizer::simplifyExternC()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "extern \"C\"")) {
if (tok->strAt(2) == "{") {
tok->linkAt(2)->deleteThis();
tok->deleteNext(2);
} else
tok->deleteNext();
tok->deleteThis();
}
}
}
void Tokenizer::simplifyRoundCurlyParentheses()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (Token::Match(tok, "[;{}:] ( {") &&
Token::simpleMatch(tok->linkAt(2), "} ) ;")) {
if (tok->str() == ":" && !Token::Match(tok->tokAt(-2),"[;{}] %type% :"))
break;
Token *end = tok->linkAt(2)->tokAt(-3);
if (Token::Match(end, "[;{}] %num%|%str% ;"))
end->deleteNext(2);
tok->linkAt(2)->previous()->deleteNext(3);
tok->deleteNext(2);
}
if (Token::Match(tok, "( { %bool%|%char%|%num%|%str%|%name% ; } )")) {
tok->deleteNext();
tok->deleteThis();
tok->deleteNext(3);
}
}
}
void Tokenizer::simplifySQL()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::simpleMatch(tok, "__CPPCHECK_EMBEDDED_SQL_EXEC__ SQL"))
continue;
const Token *end = findSQLBlockEnd(tok);
if (end == nullptr)
syntaxError(nullptr);
const std::string instruction = tok->stringifyList(end);
// delete all tokens until the embedded SQL block end
Token::eraseTokens(tok, end);
// insert "asm ( "instruction" ) ;"
tok->str("asm");
// it can happen that 'end' is NULL when wrong code is inserted
if (!tok->next())
tok->insertToken(";");
tok->insertToken(")");
tok->insertToken("\"" + instruction + "\"");
tok->insertToken("(");
// jump to ';' and continue
tok = tok->tokAt(3);
}
}
void Tokenizer::simplifyArrayAccessSyntax()
{
// 0[a] -> a[0]
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->isNumber() && Token::Match(tok, "%num% [ %name% ]")) {
const std::string number(tok->str());
Token* indexTok = tok->tokAt(2);
tok->str(indexTok->str());
tok->varId(indexTok->varId());
indexTok->str(number);
}
}
}
void Tokenizer::simplifyParameterVoid()
{
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%name% ( void )"))
tok->next()->deleteNext();
}
}
void Tokenizer::simplifyRedundantConsecutiveBraces()
{
// Remove redundant consecutive braces, i.e. '.. { { .. } } ..' -> '.. { .. } ..'.
for (Token *tok = list.front(); tok;) {
if (Token::simpleMatch(tok, "= {")) {
tok = tok->linkAt(1);
} else if (Token::simpleMatch(tok, "{ {") && Token::simpleMatch(tok->next()->link(), "} }")) {
//remove internal parentheses
tok->next()->link()->deleteThis();
tok->deleteNext();
} else
tok = tok->next();
}
}
void Tokenizer::simplifyDoublePlusAndDoubleMinus()
{
// Convert - - into + and + - into -
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (tok->next()) {
if (tok->str() == "+") {
if (tok->next()->str()[0] == '-') {
tok = tok->next();
if (tok->str().size() == 1) {
tok = tok->previous();
tok->str("-");
tok->deleteNext();
} else if (tok->isNumber()) {
tok->str(tok->str().substr(1));
tok = tok->previous();
tok->str("-");
}
continue;
}
} else if (tok->str() == "-") {
if (tok->next()->str()[0] == '-') {
tok = tok->next();
if (tok->str().size() == 1) {
tok = tok->previous();
tok->str("+");
tok->deleteNext();
} else if (tok->isNumber()) {
tok->str(tok->str().substr(1));
tok = tok->previous();
tok->str("+");
}
continue;
}
}
break;
}
}
}
/** Specify array size if it hasn't been given */
void Tokenizer::arraySize()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName() || !Token::Match(tok, "%var% [ ] ="))
continue;
bool addlength = false;
if (Token::Match(tok, "%var% [ ] = { %str% } ;")) {
Token *t = tok->tokAt(3);
t->deleteNext();
t->next()->deleteNext();
addlength = true;
}
if (addlength || Token::Match(tok, "%var% [ ] = %str% ;")) {
tok = tok->next();
const std::size_t sz = Token::getStrSize(tok->tokAt(3));
tok->insertToken(MathLib::toString(sz));
tok = tok->tokAt(5);
}
else if (Token::Match(tok, "%var% [ ] = {")) {
MathLib::biguint sz = 1;
tok = tok->next();
Token *end = tok->linkAt(3);
for (Token *tok2 = tok->tokAt(4); tok2 && tok2 != end; tok2 = tok2->next()) {
if (tok2->link() && Token::Match(tok2, "{|(|[|<")) {
if (tok2->str() == "[" && tok2->link()->strAt(1) == "=") { // designated initializer
if (Token::Match(tok2, "[ %num% ]"))
sz = std::max(sz, MathLib::toULongNumber(tok2->strAt(1)) + 1U);
else {
sz = 0;
break;
}
}
tok2 = tok2->link();
} else if (tok2->str() == ",") {
if (!Token::Match(tok2->next(), "[},]"))
++sz;
else {
tok2 = tok2->previous();
tok2->deleteNext();
}
}
}
if (sz != 0)
tok->insertToken(MathLib::toString(sz));
tok = end->next() ? end->next() : end;
}
}
}
static Token *skipTernaryOp(Token *tok)
{
unsigned int colonLevel = 1;
while (nullptr != (tok = tok->next())) {
if (tok->str() == "?") {
++colonLevel;
} else if (tok->str() == ":") {
--colonLevel;
if (colonLevel == 0) {
tok = tok->next();
break;
}
}
if (tok->link() && Token::Match(tok, "[(<]"))
tok = tok->link();
else if (Token::Match(tok->next(), "[{};)]"))
break;
}
if (colonLevel > 0) // Ticket #5214: Make sure the ':' matches the proper '?'
return nullptr;
return tok;
}
const Token * Tokenizer::startOfExecutableScope(const Token * tok)
{
if (tok->str() != ")")
return nullptr;
tok = isFunctionHead(tok, ":{", true);
if (Token::Match(tok, ": %name% [({]")) {
while (Token::Match(tok, "[:,] %name% [({]"))
tok = tok->linkAt(2)->next();
}
return (tok && tok->str() == "{") ? tok : nullptr;
}
/** simplify labels and case|default in the code: add a ";" if not already in.*/
void Tokenizer::simplifyLabelsCaseDefault()
{
const bool cpp = isCPP();
bool executablescope = false;
unsigned int indentLevel = 0;
for (Token *tok = list.front(); tok; tok = tok->next()) {
// Simplify labels in the executable scope..
Token *start = const_cast<Token *>(startOfExecutableScope(tok));
if (start) {
tok = start;
executablescope = true;
}
if (!executablescope)
continue;
if (tok->str() == "{") {
if (tok->previous()->str() == "=")
tok = tok->link();
else
++indentLevel;
} else if (tok->str() == "}") {
--indentLevel;
if (indentLevel == 0) {
executablescope = false;
continue;
}
} else if (Token::Match(tok, "(|["))
tok = tok->link();
if (Token::Match(tok, "[;{}:] case")) {
while (nullptr != (tok = tok->next())) {
if (Token::Match(tok, "(|[")) {
tok = tok->link();
} else if (tok->str() == "?") {
Token *tok1 = skipTernaryOp(tok);
if (!tok1) {
syntaxError(tok);
}
tok = tok1;
}
if (Token::Match(tok->next(),"[:{};]"))
break;
}
if (!tok)
break;
if (tok->str() != "case" && tok->next() && tok->next()->str() == ":") {
tok = tok->next();
if (!tok->next())
syntaxError(tok);
if (tok->next()->str() != ";" && tok->next()->str() != "case")
tok->insertToken(";");
else
tok = tok->previous();
} else {
syntaxError(tok);
}
} else if (Token::Match(tok, "[;{}] %name% : !!;")) {
if (!cpp || !Token::Match(tok->next(), "class|struct|enum")) {
tok = tok->tokAt(2);
tok->insertToken(";");
}
}
}
}
void Tokenizer::simplifyCaseRange()
{
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "case %num% . . . %num% :")) {
const MathLib::bigint start = MathLib::toLongNumber(tok->strAt(1));
MathLib::bigint end = MathLib::toLongNumber(tok->strAt(5));
end = std::min(start + 50, end); // Simplify it 50 times at maximum
if (start < end) {
tok = tok->tokAt(2);
tok->str(":");
tok->deleteNext();
tok->next()->str("case");
for (MathLib::bigint i = end-1; i > start; i--) {
tok->insertToken(":");
tok->insertToken(MathLib::toString(i));
tok->insertToken("case");
}
}
} else if (Token::Match(tok, "case %char% . . . %char% :")) {
const char start = tok->strAt(1)[1];
const char end = tok->strAt(5)[1];
if (start < end) {
tok = tok->tokAt(2);
tok->str(":");
tok->deleteNext();
tok->next()->str("case");
for (char i = end - 1; i > start; i--) {
tok->insertToken(":");
tok->insertToken(std::string(1, '\'') + i + '\'');
tok->insertToken("case");
}
}
}
}
}
void Tokenizer::simplifyTemplates()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
// Ticket #6181: normalize C++11 template parameter list closing syntax
if (tok->str() == "<" && mTemplateSimplifier->templateParameters(tok)) {
Token *endTok = tok->findClosingBracket();
if (endTok && endTok->str() == ">>") {
endTok->str(">");
endTok->insertToken(">");
}
}
}
mTemplateSimplifier->simplifyTemplates(
#ifdef MAXTIME
mMaxTime,
#else
0, // ignored
#endif
mCodeWithTemplates);
}
//---------------------------------------------------------------------------
static bool setVarIdParseDeclaration(const Token **tok, const std::map<std::string,unsigned int> &variableId, bool executableScope, bool cpp, bool c)
{
const Token *tok2 = *tok;
if (!tok2->isName())
return false;
unsigned int typeCount = 0;
unsigned int singleNameCount = 0;
bool hasstruct = false; // Is there a "struct" or "class"?
bool bracket = false;
bool ref = false;
while (tok2) {
if (tok2->isName()) {
if (cpp && Token::Match(tok2, "namespace|public|private|protected"))
return false;
if (Token::Match(tok2, "struct|union|enum") || (!c && Token::Match(tok2, "class|typename"))) {
hasstruct = true;
typeCount = 0;
singleNameCount = 0;
} else if (tok2->str() == "const") {
; // just skip "const"
} else if (!hasstruct && variableId.find(tok2->str()) != variableId.end() && tok2->previous()->str() != "::") {
++typeCount;
tok2 = tok2->next();
if (!tok2 || tok2->str() != "::")
break;
} else {
if (tok2->str() != "void" || Token::Match(tok2, "void const| *|(")) // just "void" cannot be a variable type
++typeCount;
++singleNameCount;
}
} else if (!c && ((TemplateSimplifier::templateParameters(tok2) > 0) ||
Token::simpleMatch(tok2, "< >") /* Ticket #4764 */)) {
const Token *start = *tok;
if (Token::Match(start->previous(), "%or%|%oror%|&&|&|^|+|-|*|/"))
return false;
const Token * tok3 = tok2->findClosingBracket();
if (tok3 == nullptr) { /* Ticket #8151 */
throw tok2;
}
tok2 = tok3;
if (tok2->str() != ">")
break;
singleNameCount = 1;
if (Token::Match(tok2, "> %name% %or%|%oror%|&&|&|^|+|-|*|/"))
return false;
if (Token::Match(tok2, "> %name% )")) {
if (Token::Match(tok2->linkAt(2)->previous(), "if|for|while ("))
return false;
if (!Token::Match(tok2->linkAt(2)->previous(), "%name% ("))
return false;
}
} else if (Token::Match(tok2, "&|&&")) {
ref = !bracket;
} else if (singleNameCount == 1 && Token::Match(tok2, "( [*&]") && Token::Match(tok2->link()->next(), "(|[")) {
bracket = true; // Skip: Seems to be valid pointer to array or function pointer
} else if (tok2->str() == "::") {
singleNameCount = 0;
} else if (tok2->str() != "*" && tok2->str() != "::") {
break;
}
tok2 = tok2->next();
}
if (tok2) {
bool isLambdaArg = false;
{
const Token *tok3 = (*tok)->previous();
if (tok3 && tok3->str() == ",") {
while (tok3 && !Token::Match(tok3,";|(|[|{")) {
if (Token::Match(tok3, ")|]"))
tok3 = tok3->link();
tok3 = tok3->previous();
}
if (tok3 && executableScope && Token::Match(tok3->previous(), "%name% (")) {
const Token *fdecl = tok3->previous();
int count = 0;
while (Token::Match(fdecl, "%name%|*")) {
fdecl = fdecl->previous();
count++;
}
if (!Token::Match(fdecl, "[;{}] %name%") || count <= 1)
return false;
}
}
if (cpp && tok3 && Token::simpleMatch(tok3->previous(), "] (") && Token::simpleMatch(tok3->link(), ") {"))
isLambdaArg = true;
}
*tok = tok2;
// In executable scopes, references must be assigned
// Catching by reference is an exception
if (executableScope && ref && !isLambdaArg) {
if (Token::Match(tok2, "(|=|{|:"))
; // reference is assigned => ok
else if (tok2->str() != ")" || tok2->link()->strAt(-1) != "catch")
return false; // not catching by reference => not declaration
}
}
// Check if array declaration is valid (#2638)
// invalid declaration: AAA a[4] = 0;
if (typeCount >= 2 && executableScope && tok2 && tok2->str() == "[") {
const Token *tok3 = tok2->link()->next();
while (tok3 && tok3->str() == "[") {
tok3 = tok3->link()->next();
}
if (Token::Match(tok3, "= %num%"))
return false;
}
return (typeCount >= 2 && tok2 && Token::Match(tok2->tokAt(-2), "!!:: %type%"));
}
static void setVarIdStructMembers(Token **tok1,
std::map<unsigned int, std::map<std::string, unsigned int> >& structMembers,
unsigned int *varId)
{
Token *tok = *tok1;
if (Token::Match(tok, "%name% = { . %name% =")) {
const unsigned int struct_varid = tok->varId();
if (struct_varid == 0)
return;
std::map<std::string, unsigned int>& members = structMembers[struct_varid];
tok = tok->tokAt(3);
while (tok->str() != "}") {
if (Token::Match(tok, "{|[|("))
tok = tok->link();
if (Token::Match(tok->previous(), "[,{] . %name% =")) {
tok = tok->next();
const std::map<std::string, unsigned int>::iterator it = members.find(tok->str());
if (it == members.end()) {
members[tok->str()] = ++(*varId);
tok->varId(*varId);
} else {
tok->varId(it->second);
}
}
tok = tok->next();
}
return;
}
while (Token::Match(tok->next(), ")| . %name% !!(")) {
const unsigned int struct_varid = tok->varId();
tok = tok->tokAt(2);
if (struct_varid == 0)
continue;
if (tok->str() == ".")
tok = tok->next();
// Don't set varid for template function
if (TemplateSimplifier::templateParameters(tok->next()) > 0)
break;
std::map<std::string, unsigned int>& members = structMembers[struct_varid];
const std::map<std::string, unsigned int>::iterator it = members.find(tok->str());
if (it == members.end()) {
members[tok->str()] = ++(*varId);
tok->varId(*varId);
} else {
tok->varId(it->second);
}
}
// tok can't be null
*tok1 = tok;
}
void Tokenizer::setVarIdClassDeclaration(const Token * const startToken,
const VariableMap &variableMap,
const unsigned int scopeStartVarId,
std::map<unsigned int, std::map<std::string,unsigned int> >& structMembers)
{
// end of scope
const Token * const endToken = startToken->link();
// determine class name
std::string className;
for (const Token *tok = startToken->previous(); tok; tok = tok->previous()) {
if (!tok->isName() && tok->str() != ":")
break;
if (Token::Match(tok, "class|struct|enum %type% [:{]")) {
className = tok->next()->str();
break;
}
}
// replace varids..
unsigned int indentlevel = 0;
bool initList = false;
bool inEnum = false;
const Token *initListArgLastToken = nullptr;
for (Token *tok = startToken->next(); tok != endToken; tok = tok->next()) {
if (!tok)
syntaxError(nullptr);
if (initList) {
if (tok == initListArgLastToken)
initListArgLastToken = nullptr;
else if (!initListArgLastToken &&
Token::Match(tok->previous(), "%name%|>|>> {|(") &&
Token::Match(tok->link(), "}|) ,|{"))
initListArgLastToken = tok->link();
}
if (tok->str() == "{") {
inEnum = isEnumStart(tok);
if (initList && !initListArgLastToken)
initList = false;
++indentlevel;
} else if (tok->str() == "}") {
--indentlevel;
inEnum = false;
} else if (initList && indentlevel == 0 && Token::Match(tok->previous(), "[,:] %name% [({]")) {
const std::map<std::string, unsigned int>::const_iterator it = variableMap.find(tok->str());
if (it != variableMap.end()) {
tok->varId(it->second);
}
} else if (tok->isName() && tok->varId() <= scopeStartVarId) {
if (indentlevel > 0 || initList) {
if (Token::Match(tok->previous(), "::|.") && tok->strAt(-2) != "this" && !Token::simpleMatch(tok->tokAt(-5), "( * this ) ."))
continue;
if (!tok->next())
syntaxError(nullptr);
if (tok->next()->str() == "::") {
if (tok->str() == className)
tok = tok->tokAt(2);
else
continue;
}
if (!inEnum) {
const std::map<std::string, unsigned int>::const_iterator it = variableMap.find(tok->str());
if (it != variableMap.end()) {
tok->varId(it->second);
setVarIdStructMembers(&tok, structMembers, variableMap.getVarId());
}
}
}
} else if (indentlevel == 0 && tok->str() == ":" && !initListArgLastToken)
initList = true;
}
}
// Update the variable ids..
// Parse each function..
static void setVarIdClassFunction(const std::string &classname,
Token * const startToken,
const Token * const endToken,
const std::map<std::string, unsigned int> &varlist,
std::map<unsigned int, std::map<std::string, unsigned int> >& structMembers,
unsigned int *varId_)
{
for (Token *tok2 = startToken; tok2 && tok2 != endToken; tok2 = tok2->next()) {
if (tok2->varId() != 0 || !tok2->isName())
continue;
if (Token::Match(tok2->tokAt(-2), ("!!" + classname + " ::").c_str()))
continue;
if (Token::Match(tok2->tokAt(-4), "%name% :: %name% ::")) // Currently unsupported
continue;
if (Token::Match(tok2->tokAt(-2), "!!this .") && !Token::simpleMatch(tok2->tokAt(-5), "( * this ) ."))
continue;
const std::map<std::string,unsigned int>::const_iterator it = varlist.find(tok2->str());
if (it != varlist.end()) {
tok2->varId(it->second);
setVarIdStructMembers(&tok2, structMembers, varId_);
}
}
}
void Tokenizer::setVarId()
{
// Clear all variable ids
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->isName())
tok->varId(0);
}
setPodTypes();
setVarIdPass1();
setVarIdPass2();
}
// Variable declarations can't start with "return" etc.
#define NOTSTART_C "case", "default", "goto", "NOT", "return", "sizeof", "typedef"
static const std::set<std::string> notstart_c = { NOTSTART_C };
static const std::set<std::string> notstart_cpp = { NOTSTART_C,
"delete", "friend", "new", "throw", "using", "virtual", "explicit", "const_cast", "dynamic_cast", "reinterpret_cast", "static_cast", "template"
};
void Tokenizer::setVarIdPass1()
{
// Variable declarations can't start with "return" etc.
const std::set<std::string>& notstart = (isC()) ? notstart_c : notstart_cpp;
VariableMap variableMap;
std::map<unsigned int, std::map<std::string, unsigned int> > structMembers;
std::stack<VarIdScopeInfo> scopeStack;
scopeStack.push(VarIdScopeInfo());
std::stack<const Token *> functionDeclEndStack;
const Token *functionDeclEndToken = nullptr;
bool initlist = false;
bool inlineFunction = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->isOp())
continue;
if (tok == functionDeclEndToken) {
functionDeclEndStack.pop();
functionDeclEndToken = functionDeclEndStack.empty() ? nullptr : functionDeclEndStack.top();
if (tok->str() == ":")
initlist = true;
else if (tok->str() == ";") {
if (!variableMap.leaveScope())
cppcheckError(tok);
} else if (tok->str() == "{")
scopeStack.push(VarIdScopeInfo(true, scopeStack.top().isStructInit || tok->strAt(-1) == "=", /*isEnum=*/false, *variableMap.getVarId()));
} else if (!initlist && tok->str()=="(") {
const Token * newFunctionDeclEnd = nullptr;
if (!scopeStack.top().isExecutable)
newFunctionDeclEnd = isFunctionHead(tok, "{:;");
else {
Token const * const tokenLinkNext = tok->link()->next();
if (tokenLinkNext && tokenLinkNext->str() == "{") // might be for- or while-loop or if-statement
newFunctionDeclEnd = tokenLinkNext;
}
if (newFunctionDeclEnd && newFunctionDeclEnd != functionDeclEndToken) {
functionDeclEndStack.push(newFunctionDeclEnd);
functionDeclEndToken = newFunctionDeclEnd;
variableMap.enterScope();
}
} else if (Token::Match(tok, "{|}")) {
inlineFunction = false;
const Token * const startToken = (tok->str() == "{") ? tok : tok->link();
// parse anonymous unions as part of the current scope
if (!Token::Match(startToken->previous(), "union|struct|enum {") &&
!(initlist && Token::Match(startToken->previous(), "%name%|>|>>") && Token::Match(startToken->link(), "} ,|{"))) {
if (tok->str() == "{") {
bool isExecutable;
const Token *prev = tok->previous();
while (Token::Match(prev, "%name%|."))
prev = prev->previous();
const bool isLambda = prev && prev->str() == ")" && Token::simpleMatch(prev->link()->previous(), "] (");
if ((!isLambda && (tok->strAt(-1) == ")" || Token::Match(tok->tokAt(-2), ") %type%"))) ||
(initlist && tok->strAt(-1) == "}")) {
isExecutable = true;
} else {
isExecutable = ((scopeStack.top().isExecutable || initlist || tok->strAt(-1) == "else") &&
!isClassStructUnionEnumStart(tok));
if (!(scopeStack.top().isStructInit || tok->strAt(-1) == "="))
variableMap.enterScope();
}
initlist = false;
scopeStack.push(VarIdScopeInfo(isExecutable, scopeStack.top().isStructInit || tok->strAt(-1) == "=", isEnumStart(tok), *variableMap.getVarId()));
} else { /* if (tok->str() == "}") */
bool isNamespace = false;
for (const Token *tok1 = tok->link()->previous(); tok1 && tok1->isName(); tok1 = tok1->previous()) {
if (tok1->str() == "namespace") {
isNamespace = true;
break;
}
}
// Set variable ids in class declaration..
if (!initlist && !isC() && !scopeStack.top().isExecutable && tok->link() && !isNamespace) {
setVarIdClassDeclaration(tok->link(),
variableMap,
scopeStack.top().startVarid,
structMembers);
}
if (!scopeStack.top().isStructInit) {
variableMap.leaveScope();
}
scopeStack.pop();
if (scopeStack.empty()) { // should be impossible
scopeStack.push(VarIdScopeInfo());
}
}
}
}
if (!scopeStack.top().isStructInit &&
(tok == list.front() ||
Token::Match(tok, "[;{}]") ||
(tok->str() == "(" && isFunctionHead(tok,"{")) ||
(tok->str() == "(" && !scopeStack.top().isExecutable && isFunctionHead(tok,";:")) ||
(tok->str() == "," && (!scopeStack.top().isExecutable || inlineFunction)) ||
(tok->isName() && endsWith(tok->str(), ':')))) {
// No variable declarations in sizeof
if (Token::simpleMatch(tok->previous(), "sizeof (")) {
continue;
}
if (mSettings->terminated())
return;
// locate the variable name..
const Token *tok2 = (tok->isName()) ? tok : tok->next();
// private: protected: public: etc
while (tok2 && endsWith(tok2->str(), ':')) {
tok2 = tok2->next();
}
if (!tok2)
break;
// Variable declaration can't start with "return", etc
if (notstart.find(tok2->str()) != notstart.end())
continue;
if (!isC() && Token::simpleMatch(tok2, "const new"))
continue;
bool decl;
try { /* Ticket #8151 */
decl = setVarIdParseDeclaration(&tok2, variableMap.map(), scopeStack.top().isExecutable, isCPP(), isC());
} catch (const Token * errTok) {
syntaxError(errTok);
}
if (decl) {
if (tok->str() == "(" && isFunctionHead(tok,"{") && scopeStack.top().isExecutable)
inlineFunction = true;
const Token* prev2 = tok2->previous();
if (Token::Match(prev2, "%type% [;[=,)]") && tok2->previous()->str() != "const")
;
else if (Token::Match(prev2, "%type% :") && tok->strAt(-1) == "for")
;
else if (Token::Match(prev2, "%type% ( !!)") && Token::simpleMatch(tok2->link(), ") ;")) {
// In C++ , a variable can't be called operator+ or something like that.
if (isCPP() &&
prev2->isOperatorKeyword())
continue;
const Token *tok3 = tok2->next();
if (!tok3->isStandardType() && tok3->str() != "void" && !Token::Match(tok3, "struct|union|class %type%") && tok3->str() != "." && !Token::Match(tok2->link()->previous(), "[&*]")) {
if (!scopeStack.top().isExecutable) {
// Detecting initializations with () in non-executable scope is hard and often impossible to be done safely. Thus, only treat code as a variable that definitely is one.
decl = false;
bool rhs = false;
for (; tok3; tok3 = tok3->nextArgumentBeforeCreateLinks2()) {
if (tok3->str() == "=") {
rhs = true;
continue;
}
if (tok3->str() == ",") {
rhs = false;
continue;
}
if (rhs)
continue;
if (tok3->isLiteral() ||
(tok3->isName() && variableMap.hasVariable(tok3->str())) ||
tok3->isOp() ||
tok3->str() == "(" ||
notstart.find(tok3->str()) != notstart.end()) {
decl = true;
break;
}
}
}
} else
decl = false;
} else if (isCPP() && Token::Match(prev2, "%type% {") && Token::simpleMatch(tok2->link(), "} ;")) { // C++11 initialization style
if (Token::Match(prev2, "do|try|else") || Token::Match(prev2->tokAt(-2), "struct|class|:"))
continue;
} else
decl = false;
if (decl) {
variableMap.addVariable(prev2->str());
// set varid for template parameters..
tok = tok->next();
while (Token::Match(tok, "%name%|::"))
tok = tok->next();
if (tok && tok->str() == "<") {
const Token *end = tok->findClosingBracket();
while (tok != end) {
if (tok->isName()) {
const std::map<std::string, unsigned int>::const_iterator it = variableMap.find(tok->str());
if (it != variableMap.end())
tok->varId(it->second);
}
tok = tok->next();
}
}
tok = tok2->previous();
}
}
}
if (tok->isName()) {
// don't set variable id after a struct|enum|union
if (Token::Match(tok->previous(), "struct|enum|union") || (isCPP() && tok->strAt(-1) == "class"))
continue;
if (!isC()) {
if (tok->previous() && tok->previous()->str() == "::")
continue;
if (tok->next() && tok->next()->str() == "::")
continue;
}
// function declaration inside executable scope? Function declaration is of form: type name "(" args ")"
if (scopeStack.top().isExecutable && Token::Match(tok, "%name% [,)]")) {
bool par = false;
const Token *start, *end;
// search begin of function declaration
for (start = tok; Token::Match(start, "%name%|*|&|,|("); start = start->previous()) {
if (start->str() == "(") {
if (par)
break;
par = true;
}
if (Token::Match(start, "[(,]")) {
if (!Token::Match(start, "[(,] %type% %name%|*|&"))
break;
}
if (start->varId() > 0U)
break;
}
// search end of function declaration
for (end = tok->next(); Token::Match(end, "%name%|*|&|,"); end = end->next()) {}
// there are tokens which can't appear at the begin of a function declaration such as "return"
const bool isNotstartKeyword = start->next() && notstart.find(start->next()->str()) != notstart.end();
// now check if it is a function declaration
if (Token::Match(start, "[;{}] %type% %name%|*") && par && Token::simpleMatch(end, ") ;") && !isNotstartKeyword)
// function declaration => don't set varid
continue;
}
if (!scopeStack.top().isEnum) {
const std::map<std::string, unsigned int>::const_iterator it = variableMap.find(tok->str());
if (it != variableMap.end()) {
tok->varId(it->second);
setVarIdStructMembers(&tok, structMembers, variableMap.getVarId());
}
}
} else if (Token::Match(tok, "::|. %name%")) {
// Don't set varid after a :: or . token
tok = tok->next();
} else if (tok->str() == ":" && Token::Match(tok->tokAt(-2), "class %type%")) {
do {
tok = tok->next();
} while (tok && (tok->isName() || tok->str() == ","));
if (!tok)
break;
tok = tok->previous();
}
}
mVarId = *variableMap.getVarId();
}
namespace {
struct Member {
Member(const std::list<std::string> &s, const std::list<const Token *> &ns, Token *t) : usingnamespaces(ns), scope(s), tok(t) {}
std::list<const Token *> usingnamespaces;
std::list<std::string> scope;
Token *tok;
};
struct ScopeInfo2 {
ScopeInfo2(const std::string &name_, const Token *bodyEnd_) : name(name_), bodyEnd(bodyEnd_) {}
const std::string name;
const Token * const bodyEnd;
};
}
static std::string getScopeName(const std::list<ScopeInfo2> &scopeInfo)
{
std::string ret;
for (const ScopeInfo2 &si : scopeInfo)
ret += (ret.empty() ? "" : " :: ") + (si.name);
return ret;
}
static Token * matchMemberName(const std::list<std::string> &scope, const Token *nsToken, Token *memberToken, const std::list<ScopeInfo2> &scopeInfo)
{
std::list<ScopeInfo2>::const_iterator scopeIt = scopeInfo.begin();
// Current scope..
for (std::list<std::string>::const_iterator it = scope.begin(); it != scope.end(); ++it) {
if (scopeIt == scopeInfo.end() || scopeIt->name != *it)
return nullptr;
++scopeIt;
}
// using namespace..
if (nsToken) {
while (Token::Match(nsToken, "%name% ::")) {
if (scopeIt != scopeInfo.end() && nsToken->str() == scopeIt->name) {
nsToken = nsToken->tokAt(2);
++scopeIt;
} else {
return nullptr;
}
}
if (!Token::Match(nsToken, "%name% ;"))
return nullptr;
if (scopeIt == scopeInfo.end() || nsToken->str() != scopeIt->name)
return nullptr;
++scopeIt;
}
// Parse member tokens..
while (scopeIt != scopeInfo.end()) {
if (!Token::Match(memberToken, "%name% ::|<"))
return nullptr;
if (memberToken->str() != scopeIt->name)
return nullptr;
if (memberToken->next()->str() == "<") {
memberToken = memberToken->next()->findClosingBracket();
if (!Token::simpleMatch(memberToken, "> ::"))
return nullptr;
}
memberToken = memberToken->tokAt(2);
++scopeIt;
}
return Token::Match(memberToken, "~| %name%") ? memberToken : nullptr;
}
static Token * matchMemberName(const Member &member, const std::list<ScopeInfo2> &scopeInfo)
{
if (scopeInfo.empty())
return nullptr;
// Does this member match without "using namespace"..
Token *ret = matchMemberName(member.scope, nullptr, member.tok, scopeInfo);
if (ret)
return ret;
// Try to match member using the "using namespace ..." namespaces..
for (const Token *ns : member.usingnamespaces) {
ret = matchMemberName(member.scope, ns, member.tok, scopeInfo);
if (ret)
return ret;
}
return nullptr;
}
static Token * matchMemberVarName(const Member &var, const std::list<ScopeInfo2> &scopeInfo)
{
Token *tok = matchMemberName(var, scopeInfo);
return Token::Match(tok, "%name% !!(") ? tok : nullptr;
}
static Token * matchMemberFunctionName(const Member &func, const std::list<ScopeInfo2> &scopeInfo)
{
Token *tok = matchMemberName(func, scopeInfo);
return Token::Match(tok, "~| %name% (") ? tok : nullptr;
}
void Tokenizer::setVarIdPass2()
{
std::map<unsigned int, std::map<std::string, unsigned int> > structMembers;
// Member functions and variables in this source
std::list<Member> allMemberFunctions;
std::list<Member> allMemberVars;
if (!isC()) {
std::map<const Token *, std::string> endOfScope;
std::list<std::string> scope;
std::list<const Token *> usingnamespaces;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->previous() || Token::Match(tok->previous(), "[;{}]")) {
if (Token::Match(tok, "using namespace %name% ::|;")) {
Token *endtok = tok->tokAt(2);
while (Token::Match(endtok, "%name% ::"))
endtok = endtok->tokAt(2);
if (Token::Match(endtok, "%name% ;"))
usingnamespaces.push_back(tok->tokAt(2));
tok = endtok;
continue;
} else if (Token::Match(tok, "namespace %name% {")) {
scope.push_back(tok->strAt(1));
endOfScope[tok->linkAt(2)] = tok->strAt(1);
}
}
if (tok->str() == "}") {
const std::map<const Token *, std::string>::iterator it = endOfScope.find(tok);
if (it != endOfScope.end())
scope.remove(it->second);
}
Token* const tok1 = tok;
if (Token::Match(tok->previous(), "!!:: %name% :: ~| %name%"))
tok = tok->next();
else if (Token::Match(tok->previous(), "!!:: %name% <") && Token::Match(tok->next()->findClosingBracket(),"> :: ~| %name%"))
tok = tok->next()->findClosingBracket()->next();
else
continue;
while (Token::Match(tok, ":: ~| %name%")) {
tok = tok->next();
if (tok->str() == "~")
tok = tok->next();
else if (Token::Match(tok, "%name% <") && Token::Match(tok->next()->findClosingBracket(),"> :: ~| %name%"))
tok = tok->next()->findClosingBracket()->next();
else if (Token::Match(tok, "%name% ::"))
tok = tok->next();
else
break;
}
if (!tok->next())
syntaxError(tok);
if (Token::Match(tok, "%name% ("))
allMemberFunctions.emplace_back(scope, usingnamespaces, tok1);
else
allMemberVars.emplace_back(scope, usingnamespaces, tok1);
}
}
std::list<ScopeInfo2> scopeInfo;
// class members..
std::map<std::string, std::map<std::string, unsigned int> > varsByClass;
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (tok->str() == "}" && !scopeInfo.empty() && tok == scopeInfo.back().bodyEnd)
scopeInfo.pop_back();
if (!Token::Match(tok, "namespace|class|struct %name% {|:|::"))
continue;
const std::string &scopeName(getScopeName(scopeInfo));
const std::string scopeName2(scopeName.empty() ? std::string() : (scopeName + " :: "));
std::list<const Token *> classnameTokens;
classnameTokens.push_back(tok->next());
const Token* tokStart = tok->tokAt(2);
while (Token::Match(tokStart, ":: %name%")) {
classnameTokens.push_back(tokStart->next());
tokStart = tokStart->tokAt(2);
}
std::string classname;
for (const Token *it : classnameTokens)
classname += (classname.empty() ? "" : " :: ") + it->str();
std::map<std::string, unsigned int> &thisClassVars = varsByClass[scopeName2 + classname];
while (Token::Match(tokStart, ":|::|,|%name%")) {
if (Token::Match(tokStart, "%name% <")) {
tokStart = tokStart->next()->findClosingBracket();
if (tokStart)
tokStart = tokStart->next();
continue;
}
if (Token::Match(tokStart, "%name% ,|{")) {
const std::map<std::string, unsigned int>& baseClassVars = varsByClass[tokStart->str()];
thisClassVars.insert(baseClassVars.begin(), baseClassVars.end());
}
tokStart = tokStart->next();
}
if (!Token::simpleMatch(tokStart, "{"))
continue;
// What member variables are there in this class?
for (const Token *it : classnameTokens)
scopeInfo.emplace_back(it->str(), tokStart->link());
for (Token *tok2 = tokStart->next(); tok2 && tok2 != tokStart->link(); tok2 = tok2->next()) {
// skip parentheses..
if (tok2->link()) {
if (tok2->str() == "{") {
if (tok2->strAt(-1) == ")" || tok2->strAt(-2) == ")")
setVarIdClassFunction(scopeName2 + classname, tok2, tok2->link(), thisClassVars, structMembers, &mVarId);
tok2 = tok2->link();
} else if (tok2->str() == "(" && tok2->link()->strAt(1) != "(") {
tok2 = tok2->link();
// Skip initialization list
while (Token::Match(tok2, ") [:,] %name% ("))
tok2 = tok2->linkAt(3);
}
}
// Found a member variable..
else if (tok2->varId() > 0)
thisClassVars[tok2->str()] = tok2->varId();
}
// Are there any member variables in this class?
if (thisClassVars.empty())
continue;
// Member variables
for (const Member &var : allMemberVars) {
Token *tok2 = matchMemberVarName(var, scopeInfo);
if (!tok2)
continue;
tok2->varId(thisClassVars[tok2->str()]);
}
if (isC() || tok->str() == "namespace")
continue;
// Set variable ids in member functions for this class..
for (const Member &func : allMemberFunctions) {
Token *tok2 = matchMemberFunctionName(func, scopeInfo);
if (!tok2)
continue;
if (tok2->str() == "~")
tok2 = tok2->linkAt(2);
else
tok2 = tok2->linkAt(1);
// If this is a function implementation.. add it to funclist
Token * start = const_cast<Token *>(isFunctionHead(tok2, "{"));
if (start) {
setVarIdClassFunction(classname, start, start->link(), thisClassVars, structMembers, &mVarId);
}
if (Token::Match(tok2, ") %name% ("))
tok2 = tok2->linkAt(2);
// constructor with initializer list
if (!Token::Match(tok2, ") : ::| %name%"))
continue;
Token *tok3 = tok2;
while (Token::Match(tok3, "[)}] [,:]")) {
tok3 = tok3->tokAt(2);
if (Token::Match(tok3, ":: %name%"))
tok3 = tok3->next();
while (Token::Match(tok3, "%name% :: %name%"))
tok3 = tok3->tokAt(2);
if (!Token::Match(tok3, "%name% (|{|<"))
break;
// set varid
const std::map<std::string, unsigned int>::const_iterator varpos = thisClassVars.find(tok3->str());
if (varpos != thisClassVars.end())
tok3->varId(varpos->second);
// goto end of var
if (tok3->strAt(1) == "<") {
tok3 = tok3->next()->findClosingBracket();
if (tok3 && tok3->next() && tok3->next()->link())
tok3 = tok3->next()->link();
} else
tok3 = tok3->linkAt(1);
}
if (Token::Match(tok3, ")|} {")) {
setVarIdClassFunction(classname, tok2, tok3->next()->link(), thisClassVars, structMembers, &mVarId);
}
}
}
}
static void linkBrackets(const Tokenizer * const tokenizer, std::stack<const Token*>& type, std::stack<Token*>& links, Token * const token, const char open, const char close)
{
if (token->str()[0] == open) {
links.push(token);
type.push(token);
} else if (token->str()[0] == close) {
if (links.empty()) {
// Error, { and } don't match.
tokenizer->unmatchedToken(token);
}
if (type.top()->str()[0] != open) {
tokenizer->unmatchedToken(type.top());
}
type.pop();
Token::createMutualLinks(links.top(), token);
links.pop();
}
}
void Tokenizer::createLinks()
{
std::stack<const Token*> type;
std::stack<Token*> links1;
std::stack<Token*> links2;
std::stack<Token*> links3;
for (Token *token = list.front(); token; token = token->next()) {
if (token->link()) {
token->link(nullptr);
}
linkBrackets(this, type, links1, token, '{', '}');
linkBrackets(this, type, links2, token, '(', ')');
linkBrackets(this, type, links3, token, '[', ']');
}
if (!links1.empty()) {
// Error, { and } don't match.
unmatchedToken(links1.top());
}
if (!links2.empty()) {
// Error, ( and ) don't match.
unmatchedToken(links2.top());
}
if (!links3.empty()) {
// Error, [ and ] don't match.
unmatchedToken(links3.top());
}
}
void Tokenizer::createLinks2()
{
if (isC())
return;
const Token * templateToken = nullptr;
bool isStruct = false;
std::stack<Token*> type;
for (Token *token = list.front(); token; token = token->next()) {
if (Token::Match(token, "%name%|> %name% [:<]"))
isStruct = true;
else if (Token::Match(token, "[;{}]"))
isStruct = false;
if (token->link()) {
if (Token::Match(token, "{|[|("))
type.push(token);
else if (!type.empty() && Token::Match(token, "}|]|)")) {
while (type.top()->str() == "<") {
if (templateToken && templateToken->next() == type.top())
templateToken = nullptr;
type.pop();
}
type.pop();
} else
token->link(nullptr);
} else if (!templateToken && !isStruct && Token::Match(token, "%oror%|&&|;")) {
if (Token::Match(token, "&& [,>]"))
continue;
// If there is some such code: A<B||C>..
// Then this is probably a template instantiation if either "B" or "C" has comparisons
if (token->tokType() == Token::eLogicalOp && !type.empty() && type.top()->str() == "<") {
const Token *prev = token->previous();
bool foundComparison = false;
while (Token::Match(prev, "%name%|%num%|%str%|%cop%|)|]") && prev != type.top()) {
if (prev->str() == ")" || prev->str() == "]")
prev = prev->link();
else if (prev->tokType() == Token::eLogicalOp)
break;
else if (prev->isComparisonOp())
foundComparison = true;
prev = prev->previous();
}
if (prev == type.top() && foundComparison)
continue;
const Token *next = token->next();
foundComparison = false;
while (Token::Match(next, "%name%|%num%|%str%|%cop%|(|[") && next->str() != ">") {
if (next->str() == "(" || next->str() == "[")
next = next->link();
else if (next->tokType() == Token::eLogicalOp)
break;
else if (next->isComparisonOp())
foundComparison = true;
next = next->next();
}
if (next && next->str() == ">" && foundComparison)
continue;
}
while (!type.empty() && type.top()->str() == "<")
type.pop();
} else if (token->str() == "<" && token->previous() && token->previous()->isName() && !token->previous()->varId()) {
type.push(token);
if (!templateToken && (token->previous()->str() == "template"))
templateToken = token;
} else if (token->str() == ">" || token->str() == ">>") {
if (type.empty() || type.top()->str() != "<") // < and > don't match.
continue;
Token * const top1 = type.top();
type.pop();
Token * const top2 = type.empty() ? nullptr : type.top();
type.push(top1);
if (!top2 || top2->str() != "<") {
if (token->str() == ">>")
continue;
if (token->next() &&
!Token::Match(token->next(), "%name%|>|&|&&|*|::|,|(|)|{|}|;|[|:") &&
!Token::simpleMatch(token->next(), ". . .") &&
!Token::Match(token->next(), "&& %name% ="))
continue;
}
// if > is followed by [ .. "new a<b>[" is expected
if (token->strAt(1) == "[") {
Token *prev = type.top()->previous();
while (prev && Token::Match(prev->previous(), ":: %name%"))
prev = prev->tokAt(-2);
if (prev && prev->str() != "new")
prev = prev->previous();
if (!prev || prev->str() != "new")
continue;
}
if (token->str() == ">>") {
type.pop();
type.pop();
Token::createMutualLinks(top2, token);
if (top1 == templateToken || top2 == templateToken)
templateToken = nullptr;
} else {
type.pop();
Token::createMutualLinks(top1, token);
if (top1 == templateToken)
templateToken = nullptr;
}
}
}
}
void Tokenizer::sizeofAddParentheses()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "sizeof !!("))
continue;
if (tok->next()->isLiteral() || Token::Match(tok->next(), "%name%|*|~|!|&")) {
Token *endToken = tok->next();
while (Token::simpleMatch(endToken, "* *"))
endToken = endToken->next();
while (Token::Match(endToken->next(), "%name%|%num%|%str%|[|(|.|::|++|--|!|~") || (Token::Match(endToken, "%type% * %op%|?|:|const|;|,"))) {
if (Token::Match(endToken->next(), "(|["))
endToken = endToken->linkAt(1);
else
endToken = endToken->next();
}
// Add ( after sizeof and ) behind endToken
tok->insertToken("(");
endToken->insertToken(")");
Token::createMutualLinks(tok->next(), endToken->next());
}
}
}
bool Tokenizer::simplifySizeof()
{
// Locate variable declarations and calculate the size
std::map<unsigned int, unsigned int> sizeOfVar;
std::map<unsigned int, const Token *> declTokOfVar;
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->varId() != 0 && sizeOfVar.find(tok->varId()) == sizeOfVar.end()) {
const unsigned int varId = tok->varId();
if (Token::Match(tok->tokAt(-3), "[;{}(,] %type% * %name% [;,)]") ||
Token::Match(tok->tokAt(-4), "[;{}(,] const %type% * %name% [;),]") ||
Token::Match(tok->tokAt(-2), "[;{}(,] %type% %name% [;),]") ||
Token::Match(tok->tokAt(-3), "[;{}(,] const %type% %name% [;),]")) {
const unsigned int size = sizeOfType(tok->previous());
if (size == 0) {
continue;
}
sizeOfVar[varId] = size;
declTokOfVar[varId] = tok;
}
else if (Token::Match(tok->previous(), "%type% %name% [ %num% ] [[;=]") ||
Token::Match(tok->tokAt(-2), "%type% * %name% [ %num% ] [[;=]")) {
unsigned int size = sizeOfType(tok->previous());
if (size == 0)
continue;
const Token* tok2 = tok->next();
do {
const MathLib::bigint num = MathLib::toLongNumber(tok2->strAt(1));
if (num<0)
break;
size *= (unsigned)num;
tok2 = tok2->tokAt(3);
} while (Token::Match(tok2, "[ %num% ]"));
if (Token::Match(tok2, "[;=]")) {
sizeOfVar[varId] = size;
declTokOfVar[varId] = tok;
}
if (!tok2) {
syntaxError(tok);
}
tok = tok2;
}
else if (Token::Match(tok->previous(), "%type% %name% [ %num% ] [,)]") ||
Token::Match(tok->tokAt(-2), "%type% * %name% [ %num% ] [,)]")) {
Token tempTok;
tempTok.str("*");
sizeOfVar[varId] = sizeOfType(&tempTok);
declTokOfVar[varId] = tok;
}
}
}
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() != "sizeof")
continue;
if (Token::simpleMatch(tok->next(), ". . .")) {
tok->deleteNext(3);
}
// sizeof('x')
if (Token::Match(tok->next(), "( %char% )")) {
tok->deleteNext();
tok->deleteThis();
tok->deleteNext();
std::ostringstream sz;
sz << ((isC()) ? mSettings->sizeof_int : 1);
tok->str(sz.str());
ret = true;
continue;
}
// sizeof ("text")
if (Token::Match(tok->next(), "( %str% )")) {
tok->deleteNext();
tok->deleteThis();
tok->deleteNext();
std::ostringstream ostr;
ostr << (Token::getStrLength(tok) + 1);
tok->str(ostr.str());
ret = true;
continue;
}
// sizeof(type *) => sizeof(*)
if (Token::Match(tok->next(), "( %type% * )")) {
tok->next()->deleteNext();
}
if (Token::simpleMatch(tok->next(), "( * )")) {
tok->str(MathLib::toString(sizeOfType(tok->tokAt(2))));
tok->deleteNext(3);
ret = true;
}
// sizeof( a )
else if (Token::Match(tok->next(), "( %var% )")) {
const std::map<unsigned int, unsigned int>::const_iterator sizeOfVarPos = sizeOfVar.find(tok->tokAt(2)->varId());
if (sizeOfVarPos != sizeOfVar.end()) {
tok->deleteNext();
tok->deleteThis();
tok->deleteNext();
tok->str(MathLib::toString(sizeOfVarPos->second));
ret = true;
} else {
// don't try to replace size of variable if variable has
// similar name with type (#329)
}
}
else if (Token::Match(tok->next(), "( %type% )")) {
const unsigned int size = sizeOfType(tok->tokAt(2));
if (size > 0) {
tok->str(MathLib::toString(size));
tok->deleteNext(3);
ret = true;
}
}
else if (Token::simpleMatch(tok->next(), "( *") || Token::Match(tok->next(), "( %name% [")) {
unsigned int derefs = 0;
const Token* nametok = tok->tokAt(2);
if (nametok->str() == "*") {
do {
nametok = nametok->next();
derefs++;
} while (nametok && nametok->str() == "*");
if (!Token::Match(nametok, "%name% )"))
continue;
} else {
const Token* tok2 = nametok->next();
do {
tok2 = tok2->link()->next();
derefs++;
} while (tok2 && tok2->str() == "[");
if (!tok2 || tok2->str() != ")")
continue;
}
// Some default value
MathLib::biguint size = 0;
const unsigned int varid = nametok->varId();
if (derefs != 0 && varid != 0 && declTokOfVar.find(varid) != declTokOfVar.end()) {
// Try to locate variable declaration..
const Token *decltok = declTokOfVar[varid];
if (Token::Match(decltok->previous(), "%type%|* %name% [")) {
size = sizeOfType(decltok->previous());
} else if (Token::Match(decltok->tokAt(-2), "%type% * %name%")) {
size = sizeOfType(decltok->tokAt(-2));
}
// Multi-dimensional array..
if (Token::Match(decltok, "%name% [") && Token::simpleMatch(decltok->linkAt(1), "] [")) {
const Token *tok2 = decltok;
for (unsigned int i = 0; i < derefs; i++)
tok2 = tok2->linkAt(1); // Skip all dimensions that are dereferenced before the sizeof call
while (Token::Match(tok2, "] [ %num% ]")) {
size *= MathLib::toULongNumber(tok2->strAt(2));
tok2 = tok2->linkAt(1);
}
if (Token::simpleMatch(tok2, "] ["))
continue;
}
} else if (nametok->strAt(1) == "[" && nametok->isStandardType()) {
size = sizeOfType(nametok);
if (size == 0)
continue;
const Token *tok2 = nametok->next();
while (Token::Match(tok2, "[ %num% ]")) {
size *= MathLib::toULongNumber(tok2->strAt(1));
tok2 = tok2->link()->next();
}
if (!tok2 || tok2->str() != ")")
continue;
}
if (size > 0) {
tok->str(MathLib::toString(size));
Token::eraseTokens(tok, tok->next()->link()->next());
ret = true;
}
}
}
return ret;
}
bool Tokenizer::simplifyTokenList1(const char FileName[])
{
if (mSettings->terminated())
return false;
// if MACRO
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "if|for|while|BOOST_FOREACH %name% (")) {
if (Token::simpleMatch(tok, "for each")) {
// 'for each ( )' -> 'asm ( )'
tok->str("asm");
tok->deleteNext();
} else if (tok->strAt(1) == "constexpr") {
tok->deleteNext();
} else {
syntaxError(tok);
}
}
}
// Is there C++ code in C file?
validateC();
// remove MACRO in variable declaration: MACRO int x;
removeMacroInVarDecl();
// Combine strings and character literals, e.g. L"string", L'c', "string1" "string2"
combineStringAndCharLiterals();
// replace inline SQL with "asm()" (Oracle PRO*C). Ticket: #1959
simplifySQL();
createLinks();
simplifyHeaders();
// Remove __asm..
simplifyAsm();
// Bail out if code is garbage
if (mTimerResults) {
Timer t("Tokenizer::tokenize::findGarbageCode", mSettings->showtime, mTimerResults);
findGarbageCode();
} else {
findGarbageCode();
}
checkConfiguration();
// if (x) MACRO() ..
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "if (")) {
tok = tok->next()->link();
if (Token::Match(tok, ") %name% (") &&
tok->next()->isUpperCaseName() &&
Token::Match(tok->linkAt(2), ") {|else")) {
syntaxError(tok->next());
}
}
}
if (mSettings->terminated())
return false;
// convert C++17 style nested namespaces to old style namespaces
simplifyNestedNamespace();
// simplify namespace aliases
simplifyNamespaceAliases();
// Remove [[attribute]]
simplifyCPPAttribute();
// remove __attribute__((?))
simplifyAttribute();
// Combine tokens..
combineOperators();
// Simplify the C alternative tokens (and, or, etc.)
simplifyCAlternativeTokens();
// replace 'sin(0)' to '0' and other similar math expressions
simplifyMathExpressions();
// combine "- %num%"
concatenateNegativeNumberAndAnyPositive();
// remove extern "C" and extern "C" {}
if (isCPP())
simplifyExternC();
// simplify weird but legal code: "[;{}] ( { code; } ) ;"->"[;{}] code;"
simplifyRoundCurlyParentheses();
// check for simple syntax errors..
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "> struct {") &&
Token::simpleMatch(tok->linkAt(2), "} ;")) {
syntaxError(tok);
}
}
if (!simplifyAddBraces())
return false;
sizeofAddParentheses();
// Simplify: 0[foo] -> *(foo)
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "0 [") && tok->linkAt(1)) {
tok->str("*");
tok->next()->str("(");
tok->linkAt(1)->str(")");
}
}
if (mSettings->terminated())
return false;
// Remove "inline", "register", and "restrict"
simplifyKeyword();
// simplify simple calculations inside <..>
if (isCPP()) {
Token *lt = nullptr;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "[;{}]"))
lt = nullptr;
else if (Token::Match(tok, "%type% <"))
lt = tok->next();
else if (lt && Token::Match(tok, ">|>> %name%|::|(")) {
const Token * const end = tok;
for (tok = lt; tok != end; tok = tok->next()) {
if (tok->isNumber())
mTemplateSimplifier->simplifyNumericCalculations(tok);
}
lt = tok->next();
}
}
}
// Convert K&R function declarations to modern C
simplifyVarDecl(true);
simplifyFunctionParameters();
// simplify case ranges (gcc extension)
simplifyCaseRange();
// simplify labels and 'case|default'-like syntaxes
simplifyLabelsCaseDefault();
// simplify '[;{}] * & ( %any% ) =' to '%any% ='
simplifyMulAndParens();
if (!isC() && !mSettings->library.markupFile(FileName)) {
findComplicatedSyntaxErrorsInTemplates();
}
if (mSettings->terminated())
return false;
// remove calling conventions __cdecl, __stdcall..
simplifyCallingConvention();
// Remove __declspec()
simplifyDeclspec();
validate();
// remove some unhandled macros in global scope
removeMacrosInGlobalScope();
// remove undefined macro in class definition:
// class DLLEXPORT Fred { };
// class Fred FINAL : Base { };
removeMacroInClassDef();
// That call here fixes #7190
validate();
// remove unnecessary member qualification..
removeUnnecessaryQualification();
// convert Microsoft memory functions
simplifyMicrosoftMemoryFunctions();
// convert Microsoft string functions
simplifyMicrosoftStringFunctions();
if (mSettings->terminated())
return false;
// Remove Qt signals and slots
simplifyQtSignalsSlots();
// remove Borland stuff..
simplifyBorland();
// syntax error: enum with typedef in it
checkForEnumsWithTypedef();
// Add parentheses to ternary operator where necessary
prepareTernaryOpForAST();
// Change initialisation of variable to assignment
simplifyInitVar();
// Split up variable declarations.
simplifyVarDecl(false);
// typedef..
if (mTimerResults) {
Timer t("Tokenizer::tokenize::simplifyTypedef", mSettings->showtime, mTimerResults);
simplifyTypedef();
} else {
simplifyTypedef();
}
// using A = B;
while (simplifyUsing())
;
// Add parentheses to ternary operator where necessary
// TODO: this is only necessary if one typedef simplification had a comma and was used within ?:
// If typedef handling is refactored and moved to symboldatabase someday we can remove this
prepareTernaryOpForAST();
for (Token* tok = list.front(); tok;) {
if (Token::Match(tok, "union|struct|class union|struct|class"))
tok->deleteNext();
else
tok = tok->next();
}
// class x y {
if (isCPP() && mSettings->isEnabled(Settings::INFORMATION)) {
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "class %type% %type% [:{]")) {
unhandled_macro_class_x_y(tok);
}
}
}
// catch bad typedef canonicalization
//
// to reproduce bad typedef, download upx-ucl from:
// http://packages.debian.org/sid/upx-ucl
// analyse the file src/stub/src/i386-linux.elf.interp-main.c
validate();
// The simplify enum have inner loops
if (mSettings->terminated())
return false;
// Put ^{} statements in asm()
simplifyAsm2();
// @..
simplifyAt();
// When the assembly code has been cleaned up, no @ is allowed
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "(") {
const Token *tok1 = tok;
tok = tok->link();
if (!tok)
syntaxError(tok1);
} else if (tok->str() == "@") {
syntaxError(tok);
}
}
// Order keywords "static" and "const"
simplifyStaticConst();
// convert platform dependent types to standard types
// 32 bits: size_t -> unsigned long
// 64 bits: size_t -> unsigned long long
list.simplifyPlatformTypes();
// collapse compound standard types into a single token
// unsigned long long int => long (with _isUnsigned=true,_isLong=true)
list.simplifyStdType();
if (mSettings->terminated())
return false;
// simplify bit fields..
simplifyBitfields();
if (mSettings->terminated())
return false;
// struct simplification "struct S {} s; => struct S { } ; S s ;
simplifyStructDecl();
if (mSettings->terminated())
return false;
// x = ({ 123; }); => { x = 123; }
simplifyAssignmentBlock();
if (mSettings->terminated())
return false;
simplifyVariableMultipleAssign();
// Collapse operator name tokens into single token
// operator = => operator=
simplifyOperatorName();
// Remove redundant parentheses
simplifyRedundantParentheses();
if (!isC()) {
// Handle templates..
if (mTimerResults) {
Timer t("Tokenizer::tokenize::simplifyTemplates", mSettings->showtime, mTimerResults);
simplifyTemplates();
} else {
simplifyTemplates();
}
// The simplifyTemplates have inner loops
if (mSettings->terminated())
return false;
// sometimes the "simplifyTemplates" fail and then unsimplified
// function calls etc remain. These have the "wrong" syntax. So
// this function will just fix so that the syntax is corrected.
validate(); // #6847 - invalid code
mTemplateSimplifier->cleanupAfterSimplify();
}
// Simplify pointer to standard types (C only)
simplifyPointerToStandardType();
// simplify function pointers
simplifyFunctionPointers();
// Change initialisation of variable to assignment
simplifyInitVar();
// Split up variable declarations.
simplifyVarDecl(false);
validate(); // #6772 "segmentation fault (invalid code) in Tokenizer::setVarId"
if (mTimerResults) {
Timer t("Tokenizer::tokenize::setVarId", mSettings->showtime, mTimerResults);
setVarId();
} else {
setVarId();
}
// Link < with >
createLinks2();
// specify array size
arraySize();
// The simplify enum might have inner loops
if (mSettings->terminated())
return false;
// Add std:: in front of std classes, when using namespace std; was given
simplifyNamespaceStd();
// Change initialisation of variable to assignment
simplifyInitVar();
simplifyDoublePlusAndDoubleMinus();
simplifyArrayAccessSyntax();
Token::assignProgressValues(list.front());
removeRedundantSemicolons();
simplifyParameterVoid();
simplifyRedundantConsecutiveBraces();
simplifyEmptyNamespaces();
elseif();
SimplifyNamelessRValueReferences();
validate();
return true;
}
bool Tokenizer::simplifyTokenList2()
{
// clear the _functionList so it can't contain dead pointers
deleteSymbolDatabase();
// Clear AST,ValueFlow. These will be created again at the end of this function.
for (Token *tok = list.front(); tok; tok = tok->next()) {
tok->clearAst();
tok->clearValueFlow();
}
// Convert e.g. atol("0") into 0
simplifyMathFunctions();
// f(x=g()) => x=g(); f(x)
simplifyAssignmentInFunctionCall();
// ";a+=b;" => ";a=a+b;"
simplifyCompoundAssignment();
simplifyCharAt();
// simplify references
simplifyReference();
simplifyStd();
if (mSettings->terminated())
return false;
simplifySizeof();
simplifyUndefinedSizeArray();
simplifyCasts();
// Simplify simple calculations before replace constants, this allows the replacement of constants that are calculated
// e.g. const static int value = sizeof(X)/sizeof(Y);
simplifyCalculations();
if (mSettings->terminated())
return false;
// Replace "*(ptr + num)" => "ptr[num]"
simplifyOffsetPointerDereference();
// Replace "&str[num]" => "(str + num)"
simplifyOffsetPointerReference();
removeRedundantAssignment();
simplifyRealloc();
// Change initialisation of variable to assignment
simplifyInitVar();
// Simplify variable declarations
simplifyVarDecl(false);
simplifyErrNoInWhile();
simplifyIfAndWhileAssign();
simplifyRedundantParentheses();
simplifyNestedStrcat();
simplifyFuncInWhile();
simplifyIfAndWhileAssign();
// replace strlen(str)
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "strlen ( %str% )")) {
tok->str(MathLib::toString(Token::getStrLength(tok->tokAt(2))));
tok->deleteNext(3);
}
}
bool modified = true;
while (modified) {
if (mSettings->terminated())
return false;
modified = false;
modified |= simplifyConditions();
modified |= simplifyFunctionReturn();
modified |= simplifyKnownVariables();
modified |= simplifyStrlen();
modified |= removeRedundantConditions();
modified |= simplifyRedundantParentheses();
modified |= simplifyConstTernaryOp();
modified |= simplifyCalculations();
validate();
}
// simplify redundant loops
simplifyWhile0();
removeRedundantFor();
// Remove redundant parentheses in return..
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (Token::simpleMatch(tok, "return (")) {
Token *tok2 = tok->next()->link();
if (Token::simpleMatch(tok2, ") ;")) {
tok->deleteNext();
tok2->deleteThis();
} else {
break;
}
}
}
simplifyReturnStrncat();
removeRedundantAssignment();
simplifyComma();
removeRedundantSemicolons();
simplifyFlowControl();
simplifyRedundantConsecutiveBraces();
simplifyEmptyNamespaces();
simplifyMathFunctions();
validate();
Token::assignProgressValues(list.front());
list.createAst();
// needed for #7208 (garbage code) and #7724 (ast max depth limit)
list.validateAst();
// Create symbol database and then remove const keywords
createSymbolDatabase();
mSymbolDatabase->setValueTypeInTokenList();
ValueFlow::setValues(&list, mSymbolDatabase, mErrorLogger, mSettings);
if (mSettings->terminated())
return false;
printDebugOutput(2);
return true;
}
//---------------------------------------------------------------------------
void Tokenizer::printDebugOutput(unsigned int simplification) const
{
const bool debug = (simplification != 1U && mSettings->debugSimplified) ||
(simplification != 2U && mSettings->debugnormal);
if (debug && list.front()) {
list.front()->printOut(nullptr, list.getFiles());
if (mSettings->xml)
std::cout << "<debug>" << std::endl;
if (mSymbolDatabase) {
if (mSettings->xml)
mSymbolDatabase->printXml(std::cout);
else if (mSettings->verbose) {
mSymbolDatabase->printOut("Symbol database");
}
}
if (mSettings->verbose)
list.front()->printAst(mSettings->verbose, mSettings->xml, std::cout);
list.front()->printValueFlow(mSettings->xml, std::cout);
if (mSettings->xml)
std::cout << "</debug>" << std::endl;
}
if (mSymbolDatabase && simplification == 2U && mSettings->debugwarnings) {
printUnknownTypes();
// the typeStartToken() should come before typeEndToken()
for (const Variable *var : mSymbolDatabase->variableList()) {
if (!var)
continue;
const Token * typetok = var->typeStartToken();
while (typetok && typetok != var->typeEndToken())
typetok = typetok->next();
if (typetok != var->typeEndToken()) {
reportError(var->typeStartToken(),
Severity::debug,
"debug",
"Variable::typeStartToken() of variable '" + var->name() + "' is not located before Variable::typeEndToken(). The location of the typeStartToken() is '" + var->typeStartToken()->str() + "' at line " + MathLib::toString(var->typeStartToken()->linenr()));
}
}
}
}
void Tokenizer::dump(std::ostream &out) const
{
// Create a xml data dump.
// The idea is not that this will be readable for humans. It's a
// data dump that 3rd party tools could load and get useful info from.
// tokens..
out << " <tokenlist>" << std::endl;
for (const Token *tok = list.front(); tok; tok = tok->next()) {
out << " <token id=\"" << tok << "\" file=\"" << ErrorLogger::toxml(list.file(tok)) << "\" linenr=\"" << tok->linenr() << '\"';
out << " str=\"" << ErrorLogger::toxml(tok->str()) << '\"';
out << " scope=\"" << tok->scope() << '\"';
if (tok->isName()) {
out << " type=\"name\"";
if (tok->isUnsigned())
out << " isUnsigned=\"true\"";
else if (tok->isSigned())
out << " isSigned=\"true\"";
} else if (tok->isNumber()) {
out << " type=\"number\"";
if (MathLib::isInt(tok->str()))
out << " isInt=\"True\"";
if (MathLib::isFloat(tok->str()))
out << " isFloat=\"True\"";
} else if (tok->tokType() == Token::eString)
out << " type=\"string\" strlen=\"" << Token::getStrLength(tok) << '\"';
else if (tok->tokType() == Token::eChar)
out << " type=\"char\"";
else if (tok->isBoolean())
out << " type=\"boolean\"";
else if (tok->isOp()) {
out << " type=\"op\"";
if (tok->isArithmeticalOp())
out << " isArithmeticalOp=\"True\"";
else if (tok->isAssignmentOp())
out << " isAssignmentOp=\"True\"";
else if (tok->isComparisonOp())
out << " isComparisonOp=\"True\"";
else if (tok->tokType() == Token::eLogicalOp)
out << " isLogicalOp=\"True\"";
}
if (tok->link())
out << " link=\"" << tok->link() << '\"';
if (tok->varId() > 0U)
out << " varId=\"" << MathLib::toString(tok->varId()) << '\"';
if (tok->variable())
out << " variable=\"" << tok->variable() << '\"';
if (tok->function())
out << " function=\"" << tok->function() << '\"';
if (!tok->values().empty())
out << " values=\"" << &tok->values() << '\"';
if (tok->type())
out << " type-scope=\"" << tok->type()->classScope << '\"';
if (tok->astParent())
out << " astParent=\"" << tok->astParent() << '\"';
if (tok->astOperand1())
out << " astOperand1=\"" << tok->astOperand1() << '\"';
if (tok->astOperand2())
out << " astOperand2=\"" << tok->astOperand2() << '\"';
if (!tok->originalName().empty())
out << " originalName=\"" << tok->originalName() << '\"';
if (tok->valueType()) {
const std::string vt = tok->valueType()->dump();
if (!vt.empty())
out << ' ' << vt;
}
out << "/>" << std::endl;
}
out << " </tokenlist>" << std::endl;
mSymbolDatabase->printXml(out);
if (list.front())
list.front()->printValueFlow(true, out);
}
void Tokenizer::simplifyHeaders()
{
// TODO : can we remove anything in headers here? Like unused declarations.
// Maybe if --dump is used we want to have _everything_.
if (mSettings->checkHeaders && mSettings->checkUnusedTemplates)
// Default=full analysis. All information in the headers are kept.
return;
const bool checkHeaders = mSettings->checkHeaders;
const bool removeUnusedIncludedFunctions = !mSettings->checkHeaders;
const bool removeUnusedIncludedClasses = !mSettings->checkHeaders;
const bool removeUnusedIncludedTemplates = !mSettings->checkUnusedTemplates || !mSettings->checkHeaders;
const bool removeUnusedTemplates = !mSettings->checkUnusedTemplates;
// We want to remove selected stuff from the headers but not *everything*.
// The intention here is to not damage the analysis of the source file.
// You should get all warnings in the source file.
// TODO: Remove unused types/variables/etc in headers..
// functions and types to keep
std::set<std::string> keep;
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName())
continue;
if (checkHeaders && tok->fileIndex() != 0)
continue;
if (Token::Match(tok, "%name% (") && !Token::simpleMatch(tok->linkAt(1), ") {")) {
keep.insert(tok->str());
continue;
}
if (Token::Match(tok, "%name% %name%|::|*|&|<")) {
keep.insert(tok->str());
}
}
const std::set<std::string> functionStart{"static", "const", "unsigned", "signed", "void", "bool", "char", "short", "int", "long", "float", "*"};
for (Token *tok = list.front(); tok; tok = tok->next()) {
const bool isIncluded = (tok->fileIndex() != 0);
// Remove executable code
if (isIncluded && mSettings->checkHeaders && tok->str() == "{") {
// TODO: We probably need to keep the executable code if this function is called from the source file.
const Token *prev = tok->previous();
while (prev && prev->isName())
prev = prev->previous();
if (Token::simpleMatch(prev, ")")) {
// Replace all tokens from { to } with a ";".
Token::eraseTokens(tok,tok->link()->next());
tok->str(";");
tok->link(nullptr);
}
}
if (Token::Match(tok, "[;{}]")) {
// Remove unused function declarations
if (isIncluded && removeUnusedIncludedFunctions) {
while (1) {
Token *start = tok->next();
while (start && functionStart.find(start->str()) != functionStart.end())
start = start->next();
if (Token::Match(start, "%name% (") && Token::Match(start->linkAt(1), ") const| ;") && keep.find(start->str()) == keep.end())
Token::eraseTokens(tok, start->linkAt(1)->tokAt(2));
else
break;
}
}
if (isIncluded && removeUnusedIncludedClasses) {
if (Token::Match(tok, "[;{}] class|struct %name% [:{]") && keep.find(tok->strAt(2)) == keep.end()) {
// Remove this class/struct
const Token *endToken = tok->tokAt(3);
if (endToken->str() == ":") {
endToken = endToken->next();
while (Token::Match(endToken, "%name%|,"))
endToken = endToken->next();
}
if (endToken && endToken->str() == "{" && Token::simpleMatch(endToken->link(), "} ;"))
Token::eraseTokens(tok, endToken->link()->next());
}
}
if (removeUnusedTemplates || (isIncluded && removeUnusedIncludedTemplates)) {
if (Token::Match(tok->next(), "template < %name%")) {
const Token *tok2 = tok->tokAt(3);
while (Token::Match(tok2, "%name% %name% [,=>]") || Token::Match(tok2, "typename . . . %name% [,>]")) {
if (Token::simpleMatch(tok2, "typename . . ."))
tok2 = tok2->tokAt(5);
else
tok2 = tok2->tokAt(2);
if (Token::Match(tok2, "= %name% [,>]"))
tok2 = tok2->tokAt(2);
if (tok2->str() == ",")
tok2 = tok2->next();
}
if (Token::Match(tok2, "> class|struct %name% [;:{]") && keep.find(tok2->strAt(2)) == keep.end()) {
const Token *endToken = tok2->tokAt(3);
if (endToken->str() == ":") {
endToken = endToken->next();
while (Token::Match(endToken, "%name%|,"))
endToken = endToken->next();
}
if (endToken && endToken->str() == "{")
endToken = endToken->link()->next();
if (endToken && endToken->str() == ";")
Token::eraseTokens(tok, endToken);
} else if (Token::Match(tok2, "> %type% %name% (") && Token::simpleMatch(tok2->linkAt(3), ") {") && keep.find(tok2->strAt(2)) == keep.end()) {
const Token *endToken = tok2->linkAt(3)->linkAt(1)->next();
Token::eraseTokens(tok, endToken);
}
}
}
}
}
}
void Tokenizer::removeMacrosInGlobalScope()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "(") {
tok = tok->link();
if (Token::Match(tok, ") %type% {") &&
!Token::Match(tok->next(), "const|namespace|class|struct|union|noexcept|override|final|volatile"))
tok->deleteNext();
}
if (Token::Match(tok, "%type%") && tok->isUpperCaseName() &&
(!tok->previous() || Token::Match(tok->previous(), "[;{}]") || (tok->previous()->isName() && endsWith(tok->previous()->str(), ':')))) {
const Token *tok2 = tok->next();
if (tok2 && tok2->str() == "(")
tok2 = tok2->link()->next();
if (Token::Match(tok, "%type% (") && Token::Match(tok2, "%type% (") && !Token::Match(tok2, "noexcept|throw") && isFunctionHead(tok2->next(), ":;{"))
unknownMacroError(tok);
// remove unknown macros before namespace|class|struct|union
if (Token::Match(tok2, "namespace|class|struct|union")) {
// is there a "{" for?
const Token *tok3 = tok2;
while (tok3 && !Token::Match(tok3,"[;{}()]"))
tok3 = tok3->next();
if (tok3 && tok3->str() == "{") {
Token::eraseTokens(tok, tok2);
tok->deleteThis();
}
continue;
}
// replace unknown macros before foo(
/*
if (Token::Match(tok2, "%type% (") && isFunctionHead(tok2->next(), "{")) {
std::string typeName;
for (const Token* tok3 = tok; tok3 != tok2; tok3 = tok3->next())
typeName += tok3->str();
Token::eraseTokens(tok, tok2);
tok->str(typeName);
}
*/
// remove unknown macros before foo::foo(
if (Token::Match(tok2, "%type% :: %type%")) {
const Token *tok3 = tok2;
while (Token::Match(tok3, "%type% :: %type% ::"))
tok3 = tok3->tokAt(2);
if (Token::Match(tok3, "%type% :: %type% (") && tok3->str() == tok3->strAt(2)) {
Token::eraseTokens(tok, tok2);
tok->deleteThis();
}
continue;
}
}
// Skip executable scopes
if (tok->str() == "{") {
const Token *prev = tok->previous();
while (prev && prev->isName())
prev = prev->previous();
if (prev && prev->str() == ")")
tok = tok->link();
}
}
}
//---------------------------------------------------------------------------
void Tokenizer::removeMacroInClassDef()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "class|struct %name% %name% {|:"))
continue;
const bool nextIsUppercase = tok->next()->isUpperCaseName();
const bool afterNextIsUppercase = tok->tokAt(2)->isUpperCaseName();
if (nextIsUppercase && !afterNextIsUppercase)
tok->deleteNext();
else if (!nextIsUppercase && afterNextIsUppercase)
tok->next()->deleteNext();
}
}
//---------------------------------------------------------------------------
void Tokenizer::removeMacroInVarDecl()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "[;{}] %name% (") && tok->next()->isUpperCaseName()) {
// goto ')' parentheses
const Token *tok2 = tok;
int parlevel = 0;
while (tok2) {
if (tok2->str() == "(")
++parlevel;
else if (tok2->str() == ")") {
if (--parlevel <= 0)
break;
}
tok2 = tok2->next();
}
tok2 = tok2 ? tok2->next() : nullptr;
// check if this is a variable declaration..
const Token *tok3 = tok2;
while (tok3 && tok3->isUpperCaseName())
tok3 = tok3->next();
if (tok3 && (tok3->isStandardType() || Token::Match(tok3,"const|static|struct|union|class")))
Token::eraseTokens(tok,tok2);
}
}
}
//---------------------------------------------------------------------------
void Tokenizer::removeRedundantAssignment()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "{")
tok = tok->link();
const Token * const start = const_cast<Token *>(startOfExecutableScope(tok));
if (start) {
tok = start->previous();
// parse in this function..
std::set<unsigned int> localvars;
const Token * const end = tok->next()->link();
for (Token * tok2 = tok->next(); tok2 && tok2 != end; tok2 = tok2->next()) {
// skip local class or struct
if (Token::Match(tok2, "class|struct %type% {|:")) {
// skip to '{'
tok2 = tok2->tokAt(2);
while (tok2 && tok2->str() != "{")
tok2 = tok2->next();
if (tok2)
tok2 = tok2->link(); // skip local class or struct
else
return;
} else if (Token::Match(tok2, "[;{}] %type% * %name% ;") && tok2->next()->str() != "return") {
tok2 = tok2->tokAt(3);
localvars.insert(tok2->varId());
} else if (Token::Match(tok2, "[;{}] %type% %name% ;") && tok2->next()->isStandardType()) {
tok2 = tok2->tokAt(2);
localvars.insert(tok2->varId());
} else if (tok2->varId() &&
!Token::Match(tok2->previous(), "[;{}] %name% = %char%|%num%|%name% ;")) {
localvars.erase(tok2->varId());
}
}
localvars.erase(0);
if (!localvars.empty()) {
for (Token *tok2 = tok->next(); tok2 && tok2 != end;) {
if (Token::Match(tok2, "[;{}] %type% %name% ;") && localvars.find(tok2->tokAt(2)->varId()) != localvars.end()) {
tok2->deleteNext(3);
} else if ((Token::Match(tok2, "[;{}] %type% * %name% ;") &&
localvars.find(tok2->tokAt(3)->varId()) != localvars.end()) ||
(Token::Match(tok2, "[;{}] %name% = %any% ;") &&
localvars.find(tok2->next()->varId()) != localvars.end())) {
tok2->deleteNext(4);
} else
tok2 = tok2->next();
}
}
}
}
}
void Tokenizer::simplifyRealloc()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "(|[") ||
(tok->str() == "{" && tok->previous() && tok->previous()->str() == "="))
tok = tok->link();
else if (Token::Match(tok, "[;{}] %name% = realloc (")) {
tok = tok->tokAt(3);
if (Token::simpleMatch(tok->next(), "( 0 ,")) {
//no "x = realloc(0,);"
if (!Token::simpleMatch(tok->next()->link(), ") ;") || tok->next()->link()->previous() == tok->tokAt(3))
continue;
// delete "0 ,"
tok->next()->deleteNext(2);
// Change function name "realloc" to "malloc"
tok->str("malloc");
tok = tok->next()->link();
} else {
Token *tok2 = tok->next()->link()->tokAt(-2);
//no "x = realloc(,0);"
if (!Token::simpleMatch(tok2, ", 0 ) ;") || tok2 == tok->tokAt(2))
continue;
//remove ", 0"
tok2 = tok2->previous();
tok2->deleteNext(2);
//change "realloc" to "free"
tok->str("free");
//insert "0" after "var ="
tok = tok->previous();
tok->insertToken("0");
//move "var = 0" between "free(...)" and ";"
tok2 = tok2->next();
Token::move(tok->previous(), tok->next(), tok2);
//add missing ";" after "free(...)"
tok2->insertToken(";");
//goto before last ";" and continue
tok = tok->next();
}
}
}
}
void Tokenizer::simplifyEmptyNamespaces()
{
if (isC())
return;
bool goback = false;
for (Token *tok = list.front(); tok; tok = tok ? tok->next() : nullptr) {
if (goback) {
tok = tok->previous();
goback = false;
}
if (Token::Match(tok, "(|[|{")) {
tok = tok->link();
continue;
}
if (!Token::Match(tok, "namespace %name% {"))
continue;
if (tok->strAt(3) == "}") {
tok->deleteNext(3); // remove '%name% { }'
if (!tok->previous()) {
// remove 'namespace' or replace it with ';' if isolated
tok->deleteThis();
goback = true;
} else { // '%any% namespace %any%'
tok = tok->previous(); // goto previous token
tok->deleteNext(); // remove next token: 'namespace'
}
} else {
tok = tok->tokAt(2);
}
}
}
void Tokenizer::simplifyFlowControl()
{
for (Token *begin = list.front(); begin; begin = begin->next()) {
if (Token::Match(begin, "(|[") ||
(begin->str() == "{" && begin->previous() && begin->strAt(-1) == "="))
begin = begin->link();
//function scope
if (!Token::simpleMatch(begin, ") {") && !Token::Match(begin, ") %name% {"))
continue;
Token* end = begin->linkAt(1+(begin->next()->str() == "{" ? 0 : 1));
unsigned int indentLevel = 0;
bool stilldead = false;
for (Token *tok = begin; tok && tok != end; tok = tok->next()) {
if (Token::Match(tok, "(|[")) {
tok = tok->link();
continue;
}
if (tok->str() == "{") {
if (tok->previous() && tok->previous()->str() == "=") {
tok = tok->link();
continue;
}
++indentLevel;
} else if (tok->str() == "}") {
if (indentLevel == 0)
break;
--indentLevel;
if (stilldead) {
eraseDeadCode(tok, nullptr);
if (indentLevel == 1 || tok->next()->str() != "}" || !Token::Match(tok->next()->link()->previous(), ";|{|}|do {"))
stilldead = false;
continue;
}
}
if (indentLevel == 0)
continue;
if (Token::Match(tok,"continue|break ;")) {
tok = tok->next();
eraseDeadCode(tok, nullptr);
} else if (Token::Match(tok,"return|goto") ||
(Token::Match(tok->previous(), "[;{}] %name% (") &&
mSettings->library.isnoreturn(tok)) ||
(isCPP() && tok->str() == "throw")) {
if (tok->next()->str() == "}")
syntaxError(tok->next()); // invalid code like in #6731
//TODO: ensure that we exclude user-defined 'exit|abort|throw', except for 'noreturn'
//catch the first ';'
for (Token *tok2 = tok->next(); tok2; tok2 = tok2->next()) {
if (Token::Match(tok2, "(|[")) {
tok2 = tok2->link();
} else if (tok2->str() == ";") {
tok = tok2;
eraseDeadCode(tok, nullptr);
break;
} else if (Token::Match(tok2, "[{}]"))
break;
}
//if everything is removed, then remove also the code after an inferior scope
//only if the actual scope is not special
if (indentLevel > 1 && tok->next()->str() == "}" && Token::Match(tok->next()->link()->previous(), ";|{|}|do {"))
stilldead = true;
}
}
begin = end;
}
}
bool Tokenizer::removeRedundantConditions()
{
// Return value for function. Set to true if there are any simplifications
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "if ( %bool% ) {"))
continue;
// Find matching else
Token *elseTag = tok->linkAt(4)->next();
const bool boolValue = (tok->strAt(2) == "true");
// Handle if with else
if (Token::simpleMatch(elseTag, "else {")) {
// Handle else
if (!boolValue) {
// Convert "if( false ) {aaa;} else {bbb;}" => "{bbb;}"
//remove '(false)'
tok->deleteNext(3);
//delete dead code inside scope
eraseDeadCode(tok, elseTag);
//remove 'else'
elseTag->deleteThis();
//remove 'if'
tok->deleteThis();
} else {
// Convert "if( true ) {aaa;} else {bbb;}" => "{aaa;}"
const Token *end = elseTag->next()->link()->next();
// Remove "else { bbb; }"
elseTag = elseTag->previous();
eraseDeadCode(elseTag, end);
// Remove "if( true )"
tok->deleteNext(3);
tok->deleteThis();
}
ret = true;
}
// Handle if without else
else {
if (!boolValue) {
//remove '(false)'
tok->deleteNext(3);
//delete dead code inside scope
eraseDeadCode(tok, elseTag);
//remove 'if'
tok->deleteThis();
} else {
// convert "if( true ) {aaa;}" => "{aaa;}"
tok->deleteNext(3);
tok->deleteThis();
}
ret = true;
}
}
return ret;
}
void Tokenizer::removeRedundantFor()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "[;{}] for ( %name% = %num% ; %name% < %num% ; ++| %name% ++| ) {") ||
Token::Match(tok, "[;{}] for ( %type% %name% = %num% ; %name% < %num% ; ++| %name% ++| ) {")) {
// Same variable name..
const Token* varTok = tok->tokAt(3);
const bool type = varTok->next()->isName();
if (type)
varTok = varTok->next();
const std::string varname(varTok->str());
const unsigned int varid(varTok->varId());
if (varname != varTok->strAt(4))
continue;
const Token *vartok2 = tok->linkAt(2)->previous();
if (vartok2->str() == "++")
vartok2 = vartok2->previous();
else if (vartok2->strAt(-1) != "++")
continue;
if (varname != vartok2->str())
continue;
// Check that the difference of the numeric values is 1
const MathLib::bigint num1(MathLib::toLongNumber(varTok->strAt(2)));
const MathLib::bigint num2(MathLib::toLongNumber(varTok->strAt(6)));
if (num1 + 1 != num2)
continue;
// check how loop variable is used in loop..
bool read = false;
bool write = false;
const Token* end = tok->linkAt(2)->next()->link();
for (const Token *tok2 = tok->linkAt(2); tok2 != end; tok2 = tok2->next()) {
if (tok2->str() == varname) {
if (tok2->previous()->isArithmeticalOp() &&
tok2->next() &&
(tok2->next()->isArithmeticalOp() || tok2->next()->str() == ";")) {
read = true;
} else {
read = write = true;
break;
}
}
}
// Simplify loop if loop variable isn't written
if (!write) {
Token* bodyBegin = tok->linkAt(2)->next();
// remove "for ("
tok->deleteNext(2);
// If loop variable is read then keep assignment before
// loop body..
if (type) {
tok->insertToken("{");
Token::createMutualLinks(tok->next(), bodyBegin->link());
bodyBegin->deleteThis();
tok = tok->tokAt(6);
} else if (read) {
// goto ";"
tok = tok->tokAt(4);
} else {
// remove "x = 0 ;"
tok->deleteNext(4);
}
// remove "x < 1 ; x ++ )"
tok->deleteNext(7);
if (!type) {
// Add assignment after the loop body so the loop variable
// get the correct end value
Token *tok2 = tok->next()->link();
tok2->insertToken(";");
tok2->insertToken(MathLib::toString(num2));
tok2->insertToken("=");
tok2->insertToken(varname);
tok2->next()->varId(varid);
}
}
}
}
}
void Tokenizer::removeRedundantSemicolons()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->link() && tok->str() == "(") {
tok = tok->link();
continue;
}
for (;;) {
if (Token::simpleMatch(tok, "; ;")) {
tok->deleteNext();
} else if (Token::simpleMatch(tok, "; { ; }")) {
tok->deleteNext(3);
} else {
break;
}
}
}
}
bool Tokenizer::simplifyAddBraces()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
Token const * tokRet=simplifyAddBracesToCommand(tok);
if (!tokRet)
return false;
}
return true;
}
Token *Tokenizer::simplifyAddBracesToCommand(Token *tok)
{
Token * tokEnd=tok;
if (Token::Match(tok,"for|switch|BOOST_FOREACH")) {
tokEnd=simplifyAddBracesPair(tok,true);
} else if (tok->str()=="while") {
Token *tokPossibleDo=tok->previous();
if (Token::simpleMatch(tok->previous(), "{"))
tokPossibleDo = nullptr;
else if (Token::simpleMatch(tokPossibleDo,"}"))
tokPossibleDo = tokPossibleDo->link();
if (!tokPossibleDo || tokPossibleDo->strAt(-1) != "do")
tokEnd=simplifyAddBracesPair(tok,true);
} else if (tok->str()=="do") {
tokEnd=simplifyAddBracesPair(tok,false);
if (tokEnd!=tok) {
// walk on to next token, i.e. "while"
// such that simplifyAddBracesPair does not close other braces
// before the "while"
if (tokEnd) {
tokEnd=tokEnd->next();
if (!tokEnd || tokEnd->str()!="while") // no while
syntaxError(tok);
}
}
} else if (tok->str()=="if") {
tokEnd=simplifyAddBracesPair(tok,true);
if (!tokEnd)
return nullptr;
if (tokEnd->strAt(1) == "else") {
Token * tokEndNextNext= tokEnd->tokAt(2);
if (!tokEndNextNext || tokEndNextNext->str() == "}")
syntaxError(tokEndNextNext);
if (tokEndNextNext->str() == "if")
// do not change "else if ..." to "else { if ... }"
tokEnd=simplifyAddBracesToCommand(tokEndNextNext);
else
tokEnd=simplifyAddBracesPair(tokEnd->next(),false);
}
}
return tokEnd;
}
Token *Tokenizer::simplifyAddBracesPair(Token *tok, bool commandWithCondition)
{
Token * tokCondition=tok->next();
if (!tokCondition) // Missing condition
return tok;
Token *tokAfterCondition=tokCondition;
if (commandWithCondition) {
if (tokCondition->str()=="(")
tokAfterCondition=tokCondition->link();
else
syntaxError(tok); // Bad condition
if (!tokAfterCondition || tokAfterCondition->strAt(1) == "]")
syntaxError(tok); // Bad condition
tokAfterCondition=tokAfterCondition->next();
if (!tokAfterCondition || Token::Match(tokAfterCondition, ")|}|,")) {
// No tokens left where to add braces around
return tok;
}
}
Token * tokBracesEnd=nullptr;
if (tokAfterCondition->str()=="{") {
// already surrounded by braces
tokBracesEnd=tokAfterCondition->link();
} else if (Token::simpleMatch(tokAfterCondition, "try {") &&
Token::simpleMatch(tokAfterCondition->linkAt(1), "} catch (")) {
tokAfterCondition->previous()->insertToken("{");
Token * tokOpenBrace = tokAfterCondition->previous();
Token * tokEnd = tokAfterCondition->linkAt(1)->linkAt(2)->linkAt(1);
if (!tokEnd) {
syntaxError(tokAfterCondition);
}
tokEnd->insertToken("}");
Token * tokCloseBrace = tokEnd->next();
Token::createMutualLinks(tokOpenBrace, tokCloseBrace);
tokBracesEnd = tokCloseBrace;
} else if (Token::Match(tokAfterCondition, "%name% : {")) {
tokAfterCondition->previous()->insertToken("{");
tokAfterCondition->linkAt(2)->insertToken("}");
tokBracesEnd = tokAfterCondition->linkAt(2)->next();
Token::createMutualLinks(tokAfterCondition->previous(), tokBracesEnd);
} else {
Token * tokEnd = simplifyAddBracesToCommand(tokAfterCondition);
if (!tokEnd) // Ticket #4887
return tok;
if (tokEnd->str()!="}") {
// Token does not end with brace
// Look for ; to add own closing brace after it
while (tokEnd && !Token::Match(tokEnd, ";|)|}")) {
if (tokEnd->tokType()==Token::eBracket || tokEnd->str() == "(") {
tokEnd = tokEnd->link();
if (!tokEnd) {
// Inner bracket does not close
return tok;
}
}
tokEnd=tokEnd->next();
}
if (!tokEnd || tokEnd->str() != ";") {
// No trailing ;
return tok;
}
}
tokAfterCondition->previous()->insertToken("{");
Token * tokOpenBrace=tokAfterCondition->previous();
tokEnd->insertToken("}");
Token * TokCloseBrace=tokEnd->next();
Token::createMutualLinks(tokOpenBrace,TokCloseBrace);
tokBracesEnd=TokCloseBrace;
}
return tokBracesEnd;
}
void Tokenizer::simplifyCompoundAssignment()
{
// Simplify compound assignments:
// "a+=b" => "a = a + b"
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "[;{}] (| *| (| %name%"))
continue;
if (tok->next()->str() == "return")
continue;
// backup current token..
Token * const tok1 = tok;
if (tok->next()->str() == "*")
tok = tok->next();
if (tok->next() && tok->next()->str() == "(") {
tok = tok->next()->link()->next();
} else {
// variable..
tok = tok->tokAt(2);
while (Token::Match(tok, ". %name%") ||
Token::Match(tok, "[|(")) {
if (tok->str() == ".")
tok = tok->tokAt(2);
else {
// goto "]" or ")"
tok = tok->link();
// goto next token..
tok = tok ? tok->next() : nullptr;
}
}
}
if (!tok)
break;
// Is current token at a compound assignment: +=|-=|.. ?
const std::string &str = tok->str();
std::string op; // operator used in assignment
if (tok->isAssignmentOp() && str.size() == 2)
op = str.substr(0, 1);
else if (tok->isAssignmentOp() && str.size() == 3)
op = str.substr(0, 2);
else {
tok = tok1;
continue;
}
// Remove the whole statement if it says: "+=0;", "-=0;", "*=1;" or "/=1;"
if (Token::Match(tok, "+=|-= 0 ;") ||
Token::simpleMatch(tok, "|= 0 ;") ||
Token::Match(tok, "*=|/= 1 ;")) {
tok = tok1;
while (tok->next()->str() != ";")
tok->deleteNext();
} else {
// Enclose the rhs in parentheses..
if (!Token::Match(tok->tokAt(2), "[;)]")) {
// Only enclose rhs in parentheses if there is some operator
bool someOperator = false;
for (Token *tok2 = tok->next(); tok2; tok2 = tok2->next()) {
if (tok2->link() && Token::Match(tok2, "{|[|("))
tok2 = tok2->link();
if (Token::Match(tok2->next(), "[;)]")) {
if (someOperator) {
tok->insertToken("(");
tok2->insertToken(")");
Token::createMutualLinks(tok->next(), tok2->next());
}
break;
}
someOperator |= (tok2->isOp() || tok2->str() == "?");
}
}
// simplify the compound assignment..
tok->str("=");
tok->insertToken(op);
std::stack<Token *> tokend;
for (Token *tok2 = tok->previous(); tok2 && tok2 != tok1; tok2 = tok2->previous()) {
// Don't duplicate ++ and --. Put preincrement in lhs. Put
// postincrement in rhs.
if (tok2->tokType() == Token::eIncDecOp) {
// pre increment/decrement => don't copy
if (tok2->next()->isName()) {
continue;
}
// post increment/decrement => move from lhs to rhs
tok->insertToken(tok2->str());
tok2->deleteThis();
continue;
}
// Copy token from lhs to rhs
tok->insertToken(tok2->str());
tok->next()->varId(tok2->varId());
if (Token::Match(tok->next(), "]|)|}"))
tokend.push(tok->next());
else if (Token::Match(tok->next(), "(|[|{")) {
Token::createMutualLinks(tok->next(), tokend.top());
tokend.pop();
}
}
}
}
}
bool Tokenizer::simplifyConditions()
{
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "! %bool%|%num%")) {
tok->deleteThis();
if (Token::Match(tok, "0|false"))
tok->str("true");
else
tok->str("false");
ret = true;
}
if (Token::simpleMatch(tok, "&& true &&")) {
tok->deleteNext(2);
ret = true;
}
else if (Token::simpleMatch(tok, "|| false ||")) {
tok->deleteNext(2);
ret = true;
}
else if (Token::Match(tok, "(|&& true && true &&|)")) {
tok->deleteNext(2);
ret = true;
}
else if (Token::Match(tok, "%oror%|( false %oror% false %oror%|)")) {
tok->deleteNext(2);
ret = true;
}
else if (Token::simpleMatch(tok, "( true ||") ||
Token::simpleMatch(tok, "( false &&")) {
Token::eraseTokens(tok->next(), tok->link());
ret = true;
}
else if (Token::simpleMatch(tok, "|| true )") ||
Token::simpleMatch(tok, "&& false )")) {
tok = tok->next();
Token::eraseTokens(tok->next()->link(), tok);
ret = true;
}
else if (Token::simpleMatch(tok, "&& false &&") ||
Token::simpleMatch(tok, "|| true ||")) {
//goto '('
Token *tok2 = tok;
while (tok2->previous()) {
if (tok2->previous()->str() == ")")
tok2 = tok2->previous()->link();
else {
tok2 = tok2->previous();
if (tok2->str() == "(")
break;
}
}
if (!tok2)
continue;
//move tok to 'true|false' position
tok = tok->next();
//remove everything before 'true|false'
Token::eraseTokens(tok2, tok);
//remove everything after 'true|false'
Token::eraseTokens(tok, tok2->link());
ret = true;
}
// Change numeric constant in condition to "true" or "false"
if (Token::Match(tok, "if|while ( %num% )|%oror%|&&")) {
tok->tokAt(2)->str((tok->strAt(2) != "0") ? "true" : "false");
ret = true;
}
if (Token::Match(tok, "&&|%oror% %num% )|%oror%|&&")) {
tok->next()->str((tok->next()->str() != "0") ? "true" : "false");
ret = true;
}
// Reduce "(%num% == %num%)" => "(true)"/"(false)"
if (Token::Match(tok, "&&|%oror%|(") &&
(Token::Match(tok->next(), "%num% %any% %num%") ||
Token::Match(tok->next(), "%bool% %any% %bool%")) &&
Token::Match(tok->tokAt(4), "&&|%oror%|)|?")) {
std::string cmp = tok->strAt(2);
bool result = false;
if (tok->next()->isNumber()) {
// Compare numbers
if (cmp == "==" || cmp == "!=") {
const std::string& op1(tok->next()->str());
const std::string& op2(tok->strAt(3));
bool eq = false;
if (MathLib::isInt(op1) && MathLib::isInt(op2))
eq = (MathLib::toLongNumber(op1) == MathLib::toLongNumber(op2));
else {
eq = (op1 == op2);
// It is inconclusive whether two unequal float representations are numerically equal
if (!eq && MathLib::isFloat(op1))
cmp.clear();
}
if (cmp == "==")
result = eq;
else
result = !eq;
} else {
const double op1 = MathLib::toDoubleNumber(tok->next()->str());
const double op2 = MathLib::toDoubleNumber(tok->strAt(3));
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.clear();
}
} else {
// Compare boolean
const bool op1 = (tok->next()->str() == std::string("true"));
const bool op2 = (tok->strAt(3) == std::string("true"));
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.clear();
}
if (! cmp.empty()) {
tok = tok->next();
tok->deleteNext(2);
tok->str(result ? "true" : "false");
ret = true;
}
}
}
return ret;
}
bool Tokenizer::simplifyConstTernaryOp()
{
bool ret = false;
const Token *templateParameterEnd = nullptr; // The end of the current template parameter list, if any
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "<" && mTemplateSimplifier->templateParameters(tok))
templateParameterEnd = tok->findClosingBracket();
if (tok == templateParameterEnd)
templateParameterEnd = nullptr; // End of the current template parameter list
if (tok->str() != "?")
continue;
if (!Token::Match(tok->tokAt(-2), "<|=|,|(|[|{|}|;|case|return %bool%|%num%") &&
!Token::Match(tok->tokAt(-4), "<|=|,|(|[|{|}|;|case|return ( %bool%|%num% )"))
continue;
const int offset = (tok->previous()->str() == ")") ? 2 : 1;
if (tok->strAt(-2*offset) == "<") {
if (isC() || !mTemplateSimplifier->templateParameters(tok->tokAt(-2*offset)))
continue; // '<' is less than; the condition is not a constant
}
// Find the token ":" then go to the next token
Token *colon = skipTernaryOp(tok);
if (!colon || colon->previous()->str() != ":" || !colon->next())
continue;
//handle the GNU extension: "x ? : y" <-> "x ? x : y"
if (colon->previous() == tok->next())
tok->insertToken(tok->strAt(-offset));
// go back before the condition, if possible
tok = tok->tokAt(-2);
if (offset == 2) {
// go further back before the "("
tok = tok->tokAt(-2);
//simplify the parentheses
tok->deleteNext();
tok->next()->deleteNext();
}
if (Token::Match(tok->next(), "false|0")) {
// Use code after colon, remove code before it.
Token::eraseTokens(tok, colon);
tok = tok->next();
ret = true;
}
// The condition is true. Delete the operator after the ":"..
else {
// delete the condition token and the "?"
tok->deleteNext(2);
unsigned int ternaryOplevel = 0;
for (const Token *endTok = colon; endTok; endTok = endTok->next()) {
if (Token::Match(endTok, "(|[|{")) {
endTok = endTok->link();
}
else if (endTok->str() == "?")
++ternaryOplevel;
else if (Token::Match(endTok, ")|}|]|;|,|:|>")) {
if (endTok->str() == ":" && ternaryOplevel)
--ternaryOplevel;
else if (endTok->str() == ">" && !templateParameterEnd)
;
else {
Token::eraseTokens(colon->tokAt(-2), endTok);
ret = true;
break;
}
}
}
}
}
return ret;
}
void Tokenizer::simplifyUndefinedSizeArray()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%type%")) {
Token *tok2 = tok->next();
while (tok2 && tok2->str() == "*")
tok2 = tok2->next();
if (!Token::Match(tok2, "%name% [ ] ;|["))
continue;
tok = tok2->previous();
Token *end = tok2->next();
unsigned int count = 0;
do {
end = end->tokAt(2);
++count;
} while (Token::Match(end, "[ ] [;=[]"));
if (Token::Match(end, "[;=]")) {
do {
tok2->deleteNext(2);
tok->insertToken("*");
} while (--count);
tok = end;
} else
tok = tok->tokAt(3);
}
}
}
void Tokenizer::simplifyCasts()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
// Don't remove cast in such cases:
// *((char *)a + 1) = 0;
// Remove cast when casting a function pointer:
// (*(void (*)(char *))fp)(x);
if (!tok->isName() &&
Token::simpleMatch(tok->next(), "* (") &&
!Token::Match(tok->linkAt(2), ") %name%|&")) {
tok = tok->linkAt(2);
continue;
}
// #3935 : don't remove cast in such cases:
// ((char *)a)[1] = 0;
if (tok->str() == "(" && Token::simpleMatch(tok->link(), ") [")) {
tok = tok->link();
continue;
}
// #4164 : ((unsigned char)1) => (1)
if (Token::Match(tok->next(), "( %type% ) %num%") && tok->next()->link()->previous()->isStandardType()) {
const MathLib::bigint value = MathLib::toLongNumber(tok->next()->link()->next()->str());
unsigned int bits = mSettings->char_bit * mTypeSize[tok->next()->link()->previous()->str()];
if (!tok->tokAt(2)->isUnsigned() && bits > 0)
bits--;
if (bits < 31 && value >= 0 && value < (1LL << bits)) {
tok->linkAt(1)->next()->isCast(true);
Token::eraseTokens(tok, tok->next()->link()->next());
}
continue;
}
while ((Token::Match(tok->next(), "( %type% *| *| *|&| ) *|&| %name%") && (tok->str() != ")" || tok->tokAt(2)->isStandardType())) ||
Token::Match(tok->next(), "( const| %type% * *| *|&| ) *|&| %name%") ||
Token::Match(tok->next(), "( const| %type% %type% *| *| *|&| ) *|&| %name%") ||
(!tok->isName() && (Token::Match(tok->next(), "( %type% * *| *|&| ) (") ||
Token::Match(tok->next(), "( const| %type% %type% * *| *|&| ) (")))) {
if (tok->isName() && tok->str() != "return")
break;
if (isCPP() && tok->strAt(-1) == "operator")
break;
// Remove cast..
Token::eraseTokens(tok, tok->next()->link()->next());
// Set isCasted flag.
Token *tok2 = tok->next();
if (!Token::Match(tok2, "%name% [|."))
tok2->isCast(true);
else {
// TODO: handle more complex expressions
tok2->next()->isCast(true);
}
// Remove '* &'
if (Token::simpleMatch(tok, "* &")) {
tok->deleteNext();
tok->deleteThis();
}
if (tok->str() == ")" && tok->link()->previous()) {
// If there was another cast before this, go back
// there to check it also. e.g. "(int)(char)x"
tok = tok->link()->previous();
}
}
// Replace pointer casts of 0.. "(char *)0" => "0"
while (Token::Match(tok->next(), "( %type% %type%| * *| ) 0")) {
tok->linkAt(1)->next()->isCast(true);
Token::eraseTokens(tok, tok->next()->link()->next());
if (tok->str() == ")" && tok->link()->previous()) {
// If there was another cast before this, go back
// there to check it also. e.g. "(char*)(char*)0"
tok = tok->link()->previous();
}
}
if (Token::Match(tok->next(), "dynamic_cast|reinterpret_cast|const_cast|static_cast <")) {
Token *tok2 = tok->linkAt(2);
if (!Token::simpleMatch(tok2, "> ("))
break;
tok2->tokAt(2)->isCast(true);
Token::eraseTokens(tok, tok2->next());
}
}
}
void Tokenizer::simplifyFunctionParameters()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->link() && Token::Match(tok, "{|[|(")) {
tok = tok->link();
}
// Find the function e.g. foo( x ) or foo( x, y )
else if (Token::Match(tok, "%name% ( %name% [,)]") &&
!(tok->strAt(-1) == ":" || tok->strAt(-1) == ",")) {
// We have found old style function, now we need to change it
// First step: Get list of argument names in parentheses
std::map<std::string, Token *> argumentNames;
bool bailOut = false;
Token * tokparam = nullptr;
//take count of the function name..
const std::string& funcName(tok->str());
//floating token used to check for parameters
Token *tok1 = tok;
while (nullptr != (tok1 = tok1->tokAt(2))) {
if (!Token::Match(tok1, "%name% [,)]")) {
bailOut = true;
break;
}
//same parameters: take note of the parameter
if (argumentNames.find(tok1->str()) != argumentNames.end())
tokparam = tok1;
else if (tok1->str() != funcName)
argumentNames[tok1->str()] = tok1;
else {
if (tok1->next()->str() == ")") {
if (tok1->previous()->str() == ",") {
tok1 = tok1->tokAt(-2);
tok1->deleteNext(2);
} else {
tok1 = tok1->previous();
tok1->deleteNext();
bailOut = true;
break;
}
} else {
tok1 = tok1->tokAt(-2);
tok1->next()->deleteNext(2);
}
}
if (tok1->next()->str() == ")") {
tok1 = tok1->tokAt(2);
//expect at least a type name after round brace..
if (!tok1 || !tok1->isName())
bailOut = true;
break;
}
}
//goto '('
tok = tok->next();
if (bailOut) {
tok = tok->link();
continue;
}
tok1 = tok->link()->next();
// there should be the sequence '; {' after the round parentheses
for (const Token* tok2 = tok1; tok2; tok2 = tok2->next()) {
if (Token::simpleMatch(tok2, "; {"))
break;
else if (tok2->str() == "{") {
bailOut = true;
break;
}
}
if (bailOut) {
tok = tok->link();
continue;
}
// Last step: check out if the declarations between ')' and '{' match the parameters list
std::map<std::string, Token *> argumentNames2;
while (tok1 && tok1->str() != "{") {
if (Token::Match(tok1, "(|)")) {
bailOut = true;
break;
}
if (tok1->str() == ";") {
if (tokparam) {
syntaxError(tokparam);
}
Token *tok2 = tok1->previous();
while (tok2->str() == "]")
tok2 = tok2->link()->previous();
//it should be a name..
if (!tok2->isName()) {
bailOut = true;
break;
}
if (argumentNames2.find(tok2->str()) != argumentNames2.end()) {
//same parameter names...
syntaxError(tok1);
} else
argumentNames2[tok2->str()] = tok2;
if (argumentNames.find(tok2->str()) == argumentNames.end()) {
//non-matching parameter... bailout
bailOut = true;
break;
}
}
tok1 = tok1->next();
}
if (bailOut || !tok1) {
tok = tok->link();
continue;
}
//the two containers may not hold the same size...
//in that case, the missing parameters are defined as 'int'
if (argumentNames.size() != argumentNames2.size()) {
//move back 'tok1' to the last ';'
tok1 = tok1->previous();
for (std::map<std::string, Token *>::iterator it = argumentNames.begin(); it != argumentNames.end(); ++it) {
if (argumentNames2.find(it->first) == argumentNames2.end()) {
//add the missing parameter argument declaration
tok1->insertToken(";");
tok1->insertToken(it->first);
//register the change inside argumentNames2
argumentNames2[it->first] = tok1->next();
tok1->insertToken("int");
}
}
}
while (tok->str() != ")") {
//initialize start and end tokens to be moved
Token *declStart = argumentNames2[tok->next()->str()];
Token *declEnd = declStart;
while (declStart->previous()->str() != ";" && declStart->previous()->str() != ")")
declStart = declStart->previous();
while (declEnd->next()->str() != ";" && declEnd->next()->str() != "{")
declEnd = declEnd->next();
//remove ';' after declaration
declEnd->deleteNext();
//replace the parameter name in the parentheses with all the declaration
Token::replace(tok->next(), declStart, declEnd);
//since there are changes to tokens, put tok where tok1 is
tok = declEnd->next();
}
//goto forward and continue
tok = tok->next()->link();
}
}
}
void Tokenizer::simplifyPointerToStandardType()
{
if (!isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "& %name% [ 0 ] !!["))
continue;
if (!Token::Match(tok->previous(), "[,(=]"))
continue;
// Remove '[ 0 ]' suffix
tok->next()->eraseTokens(tok->next(), tok->tokAt(5));
// Remove '&' prefix
tok = tok->previous();
if (!tok)
break;
tok->deleteNext();
}
}
void Tokenizer::simplifyFunctionPointers()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
// #2873 - do not simplify function pointer usage here:
// (void)(xy(*p)(0));
if (Token::simpleMatch(tok, ") (")) {
tok = tok->next()->link();
continue;
}
// check for function pointer cast
if (Token::Match(tok, "( %type% %type%| *| *| ( * ) (") ||
Token::Match(tok, "static_cast < %type% %type%| *| *| ( * ) (")) {
Token *tok1 = tok;
if (isCPP() && tok1->str() == "static_cast")
tok1 = tok1->next();
tok1 = tok1->next();
if (Token::Match(tok1->next(), "%type%"))
tok1 = tok1->next();
while (tok1->next()->str() == "*")
tok1 = tok1->next();
// check that the cast ends
if (!Token::Match(tok1->linkAt(4), ") )|>"))
continue;
// ok simplify this function pointer cast to an ordinary pointer cast
tok1->deleteNext();
tok1->next()->deleteNext();
Token::eraseTokens(tok1->next(), tok1->linkAt(2)->next());
continue;
}
// check for start of statement
else if (tok->previous() && !Token::Match(tok->previous(), "{|}|;|,|(|public:|protected:|private:"))
continue;
if (Token::Match(tok, "delete|else|return|throw|typedef"))
continue;
while (Token::Match(tok, "%type%|:: %type%|::"))
tok = tok->next();
Token *tok2 = (tok && tok->isName()) ? tok->next() : nullptr;
while (Token::Match(tok2, "*|&"))
tok2 = tok2->next();
if (!tok2 || tok2->str() != "(")
continue;
while (Token::Match(tok2, "(|:: %type%"))
tok2 = tok2->tokAt(2);
if (!Token::Match(tok2, "(|:: * *| %name%"))
continue;
tok2 = tok2->tokAt(2);
if (tok2->str() == "*")
tok2 = tok2->next();
while (Token::Match(tok2, "%type%|:: %type%|::"))
tok2 = tok2->next();
if (!Token::Match(tok2, "%name% ) (") &&
!Token::Match(tok2, "%name% [ ] ) (") &&
!(Token::Match(tok2, "%name% (") && Token::simpleMatch(tok2->linkAt(1), ") ) (")))
continue;
while (tok->str() != "(")
tok = tok->next();
// check that the declaration ends
if (!tok || !tok->link() || !tok->link()->next()) {
syntaxError(nullptr);
}
Token *endTok = tok->link()->next()->link();
if (Token::simpleMatch(endTok, ") throw ("))
endTok = endTok->linkAt(2);
if (!Token::Match(endTok, ") const|volatile| const|volatile| ;|,|)|=|[|{"))
continue;
while (Token::Match(endTok->next(), "const|volatile"))
endTok->deleteNext();
// ok simplify this function pointer to an ordinary pointer
Token::eraseTokens(tok->link(), endTok->next());
tok->link()->deleteThis();
while (Token::Match(tok, "( %type% ::"))
tok->deleteNext(2);
tok->deleteThis();
}
}
bool Tokenizer::simplifyFunctionReturn()
{
std::map<std::string, const Token*> functions;
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (tok->str() == "{")
tok = tok->link();
else if (Token::Match(tok, "%name% ( ) { return %bool%|%char%|%num%|%str% ; }") && tok->strAt(-1) != "::") {
const Token* const any = tok->tokAt(5);
functions[tok->str()] = any;
tok = any;
}
}
if (functions.empty())
return false;
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "(|[|=|return|%op% %name% ( ) ;|]|)|%cop%")) {
tok = tok->next();
auto it = functions.find(tok->str());
if (it != functions.cend()) {
tok->str(it->second->str());
tok->deleteNext(2);
ret = true;
}
}
}
return ret;
}
void Tokenizer::simplifyVarDecl(const bool only_k_r_fpar)
{
simplifyVarDecl(list.front(), nullptr, only_k_r_fpar);
}
void Tokenizer::simplifyVarDecl(Token * tokBegin, const Token * const tokEnd, const bool only_k_r_fpar)
{
const bool isCPP11 = mSettings->standards.cpp >= Standards::CPP11;
// Split up variable declarations..
// "int a=4;" => "int a; a=4;"
bool finishedwithkr = true;
for (Token *tok = tokBegin; tok != tokEnd; tok = tok->next()) {
if (Token::simpleMatch(tok, "= {")) {
tok = tok->next()->link();
}
if (!tok) {
syntaxError(tokBegin);
}
if (only_k_r_fpar && finishedwithkr) {
if (Token::Match(tok, "(|[|{")) {
tok = tok->link();
if (tok->next() && Token::Match(tok, ") !!{"))
tok = tok->next();
else
continue;
} else
continue;
} else if (tok->str() == "(") {
if (isCPP()) {
for (Token * tok2 = tok; tok2 && tok2 != tok->link(); tok2 = tok2->next()) {
if (Token::Match(tok2, "[(,] [")) {
// lambda function at tok2->next()
// find start of lambda body
Token * lambdaBody = tok2;
while (lambdaBody && lambdaBody != tok2->link() && lambdaBody->str() != "{")
lambdaBody = lambdaBody->next();
if (lambdaBody && lambdaBody != tok2->link() && lambdaBody->link())
simplifyVarDecl(lambdaBody, lambdaBody->link()->next(), only_k_r_fpar);
}
}
}
tok = tok->link();
}
if (!tok)
syntaxError(nullptr); // #7043 invalid code
if (tok->previous() && !Token::Match(tok->previous(), "{|}|;|)|public:|protected:|private:"))
continue;
if (Token::simpleMatch(tok, "template <"))
continue;
Token *type0 = tok;
if (!Token::Match(type0, "::|extern| %type%"))
continue;
if (Token::Match(type0, "else|return|public:|protected:|private:"))
continue;
if (isCPP11 && type0->str() == "using")
continue;
if (isCPP() && type0->str() == "namespace")
continue;
bool isconst = false;
bool isstatic = false;
Token *tok2 = type0;
unsigned int typelen = 1;
if (Token::Match(tok2, "::|extern")) {
tok2 = tok2->next();
typelen++;
}
//check if variable is declared 'const' or 'static' or both
while (tok2) {
if (!Token::Match(tok2, "const|static") && Token::Match(tok2, "%type% const|static")) {
tok2 = tok2->next();
++typelen;
}
if (tok2->str() == "const")
isconst = true;
else if (tok2->str() == "static")
isstatic = true;
else if (Token::Match(tok2, "%type% :: %type%")) {
tok2 = tok2->next();
++typelen;
}
else
break;
if (tok2->strAt(1) == "*")
break;
if (Token::Match(tok2->next(), "& %name% ,"))
break;
tok2 = tok2->next();
++typelen;
}
// strange looking variable declaration => don't split up.
if (Token::Match(tok2, "%type% *|&| %name% , %type% *|&| %name%"))
continue;
if (Token::Match(tok2, "struct|union|class %type%")) {
tok2 = tok2->next();
++typelen;
}
// check for qualification..
if (Token::Match(tok2, ":: %type%")) {
++typelen;
tok2 = tok2->next();
}
//skip combinations of templates and namespaces
while (!isC() && (Token::Match(tok2, "%type% <") || Token::Match(tok2, "%type% ::"))) {
if (tok2->next()->str() == "<" && !mTemplateSimplifier->templateParameters(tok2->next())) {
tok2 = nullptr;
break;
}
typelen += 2;
tok2 = tok2->tokAt(2);
if (tok2 && tok2->previous()->str() == "::")
continue;
unsigned int indentlevel = 0;
unsigned int parens = 0;
for (Token *tok3 = tok2; tok3; tok3 = tok3->next()) {
++typelen;
if (!parens && tok3->str() == "<") {
++indentlevel;
} else if (!parens && tok3->str() == ">") {
if (indentlevel == 0) {
tok2 = tok3->next();
break;
}
--indentlevel;
} else if (!parens && tok3->str() == ">>") {
if (indentlevel <= 1U) {
tok2 = tok3->next();
break;
}
indentlevel -= 2;
} else if (tok3->str() == "(") {
++parens;
} else if (tok3->str() == ")") {
if (!parens) {
tok2 = nullptr;
break;
}
--parens;
} else if (tok3->str() == ";") {
break;
}
}
if (Token::Match(tok2, ":: %type%")) {
++typelen;
tok2 = tok2->next();
}
}
//pattern: "%type% *| ... *| const| %name% ,|="
if (Token::Match(tok2, "%type%") ||
(tok2 && tok2->previous() && tok2->previous()->str() == ">")) {
Token *varName = tok2;
if (!tok2->previous() || tok2->previous()->str() != ">")
varName = varName->next();
else
--typelen;
//skip all the pointer part
bool isPointerOrRef = false;
while (Token::simpleMatch(varName, "*") || Token::Match(varName, "& %name% ,")) {
isPointerOrRef = true;
varName = varName->next();
}
while (Token::Match(varName, "%type% %type%")) {
if (varName->str() != "const") {
++typelen;
}
varName = varName->next();
}
//non-VLA case
if (Token::Match(varName, "%name% ,|=")) {
if (varName->str() != "operator") {
tok2 = varName->next(); // The ',' or '=' token
if (tok2->str() == "=" && (isstatic || (isconst && !isPointerOrRef))) {
//do not split const non-pointer variables..
while (tok2 && tok2->str() != "," && tok2->str() != ";") {
if (Token::Match(tok2, "{|(|["))
tok2 = tok2->link();
if (!isC() && tok2->str() == "<" && mTemplateSimplifier->templateParameters(tok2) > 0) {
tok2 = tok2->findClosingBracket();
}
if (!tok2)
syntaxError(nullptr); // #6881 invalid code
tok2 = tok2->next();
}
if (tok2 && tok2->str() == ";")
tok2 = nullptr;
}
} else
tok2 = nullptr;
}
//VLA case
else if (Token::Match(varName, "%name% [")) {
tok2 = varName->next();
while (Token::Match(tok2->link(), "] ,|=|["))
tok2 = tok2->link()->next();
if (!Token::Match(tok2, "=|,"))
tok2 = nullptr;
if (tok2 && tok2->str() == "=") {
while (tok2 && tok2->str() != "," && tok2->str() != ";") {
if (Token::Match(tok2, "{|(|["))
tok2 = tok2->link();
tok2 = tok2->next();
}
if (tok2 && tok2->str() == ";")
tok2 = nullptr;
}
}
// brace initialization
else if (Token::Match(varName, "%name% {")) {
tok2 = varName->next();
tok2 = tok2->link();
if (tok2)
tok2 = tok2->next();
if (tok2 && tok2->str() != ",")
tok2 = nullptr;
}
// parenthesis, functions can't be declared like:
// int f1(a,b), f2(c,d);
// so if there is a comma assume this is a variable declaration
else if (Token::Match(varName, "%name% (") && Token::simpleMatch(varName->linkAt(1), ") ,")) {
tok2 = varName->linkAt(1)->next();
}
else
tok2 = nullptr;
} else {
tok2 = nullptr;
}
if (!tok2) {
if (only_k_r_fpar)
finishedwithkr = false;
continue;
}
if (tok2->str() == ",") {
tok2->str(";");
//TODO: should we have to add also template '<>' links?
list.insertTokens(tok2, type0, typelen);
}
else {
Token *eq = tok2;
while (tok2) {
if (Token::Match(tok2, "{|("))
tok2 = tok2->link();
else if (!isC() && tok2->str() == "<" && tok2->previous()->isName() && !tok2->previous()->varId())
tok2 = tok2->findClosingBracket();
else if (std::strchr(";,", tok2->str()[0])) {
// "type var =" => "type var; var ="
const Token *varTok = type0->tokAt((int)typelen);
while (Token::Match(varTok, "*|&|const"))
varTok = varTok->next();
if (!varTok)
syntaxError(tok2); // invalid code
list.insertTokens(eq, varTok, 2);
eq->str(";");
// "= x, " => "= x; type "
if (tok2->str() == ",") {
tok2->str(";");
list.insertTokens(tok2, type0, typelen);
}
break;
}
if (tok2)
tok2 = tok2->next();
}
}
finishedwithkr = (only_k_r_fpar && tok2 && tok2->strAt(1) == "{");
}
}
void Tokenizer::simplifyStaticConst()
{
// This function will simplify the token list so that the qualifiers "extern", "static"
// and "const" appear in the same order as in the array below.
const std::string qualifiers[] = {"extern", "static", "const"};
// Move 'const' before all other qualifiers and types and then
// move 'static' before all other qualifiers and types, ...
for (Token *tok = list.front(); tok; tok = tok->next()) {
bool continue2 = false;
for (size_t i = 0; i < sizeof(qualifiers)/sizeof(qualifiers[0]); i++) {
// Keep searching for a qualifier
if (!tok->next() || tok->next()->str() != qualifiers[i])
continue;
// Look backwards to find the beginning of the declaration
Token* leftTok = tok;
bool behindOther = false;
for (; leftTok; leftTok = leftTok->previous()) {
for (size_t j = 0; j <= i; j++) {
if (leftTok->str() == qualifiers[j]) {
behindOther = true;
break;
}
}
if (behindOther)
break;
if (!Token::Match(leftTok, "%type%|struct|::") ||
(isCPP() && Token::Match(leftTok, "private:|protected:|public:|operator|template"))) {
break;
}
}
// The token preceding the declaration should indicate the start of a declaration
if (leftTok == tok)
continue;
if (leftTok && !behindOther && !Token::Match(leftTok, ";|{|}|(|,|private:|protected:|public:")) {
continue2 = true;
break;
}
// Move the qualifier to the left-most position in the declaration
tok->deleteNext();
if (!leftTok) {
list.front()->insertToken(qualifiers[i], emptyString, false);
list.front()->swapWithNext();
tok = list.front();
} else if (leftTok->next()) {
leftTok->next()->insertToken(qualifiers[i], emptyString, true);
tok = leftTok->next();
} else {
leftTok->insertToken(qualifiers[i]);
tok = leftTok;
}
}
if (continue2)
continue;
}
}
void Tokenizer::simplifyIfAndWhileAssign()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok->next(), "if|while ("))
continue;
const Token* tokAt3 = tok->tokAt(3);
if (!Token::Match(tokAt3, "!| (| %name% =") &&
!Token::Match(tokAt3, "!| (| %name% . %name% =") &&
!Token::Match(tokAt3, "0 == (| %name% =") &&
!Token::Match(tokAt3, "0 == (| %name% . %name% ="))
continue;
// simplifying a "while(cond) { }" condition ?
const bool iswhile(tok->next()->str() == "while");
// simplifying a "do { } while(cond);" condition ?
const bool isDoWhile = iswhile && Token::simpleMatch(tok, "}") && Token::simpleMatch(tok->link()->previous(), "do");
Token* openBraceTok = tok->link();
// delete the "if|while"
tok->deleteNext();
// Remember if there is a "!" or not. And delete it if there are.
const bool isNot(Token::Match(tok->tokAt(2), "!|0"));
if (isNot)
tok->next()->deleteNext((tok->strAt(2) == "0") ? 2 : 1);
// Delete parentheses.. and remember how many there are with
// their links.
std::stack<Token *> braces;
while (tok->next()->str() == "(") {
braces.push(tok->next()->link());
tok->deleteNext();
}
// Skip the "%name% = ..."
Token *tok2;
for (tok2 = tok->next(); tok2; tok2 = tok2->next()) {
if (tok2->str() == "(")
tok2 = tok2->link();
else if (tok2->str() == ")")
break;
}
// Insert "; if|while ( .."
tok2 = tok2->previous();
if (tok->strAt(2) == ".") {
tok2->insertToken(tok->strAt(3));
tok2->next()->varId(tok->tokAt(3)->varId());
tok2->insertToken(".");
}
tok2->insertToken(tok->next()->str());
tok2->next()->varId(tok->next()->varId());
while (! braces.empty()) {
tok2->insertToken("(");
Token::createMutualLinks(tok2->next(), braces.top());
braces.pop();
}
if (isNot)
tok2->next()->insertToken("!");
tok2->insertToken(iswhile ? "while" : "if");
if (isDoWhile) {
tok2->insertToken("}");
Token::createMutualLinks(openBraceTok, tok2->next());
}
tok2->insertToken(";");
// delete the extra "}"
if (isDoWhile)
tok->deleteThis();
// If it's a while loop, insert the assignment in the loop
if (iswhile && !isDoWhile) {
unsigned int indentlevel = 0;
Token *tok3 = tok2;
for (; tok3; tok3 = tok3->next()) {
if (tok3->str() == "{")
++indentlevel;
else if (tok3->str() == "}") {
if (indentlevel <= 1)
break;
--indentlevel;
}
}
if (tok3 && indentlevel == 1) {
tok3 = tok3->previous();
std::stack<Token *> braces2;
for (tok2 = tok2->next(); tok2 && tok2 != tok; tok2 = tok2->previous()) {
tok3->insertToken(tok2->str());
Token *newTok = tok3->next();
newTok->varId(tok2->varId());
newTok->fileIndex(tok2->fileIndex());
newTok->linenr(tok2->linenr());
// link() new tokens manually
if (tok2->link()) {
if (Token::Match(newTok, "}|)|]|>")) {
braces2.push(newTok);
} else {
Token::createMutualLinks(newTok, braces2.top());
braces2.pop();
}
}
}
}
}
}
}
void Tokenizer::simplifyVariableMultipleAssign()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%name% = %name% = %num%|%name% ;")) {
// skip intermediate assignments
Token *tok2 = tok->previous();
while (tok2 &&
tok2->str() == "=" &&
Token::Match(tok2->previous(), "%name%")) {
tok2 = tok2->tokAt(-2);
}
if (!tok2 || tok2->str() != ";") {
continue;
}
Token *stopAt = tok->tokAt(2);
const Token *valueTok = stopAt->tokAt(2);
const std::string& value(valueTok->str());
tok2 = tok2->next();
while (tok2 != stopAt) {
tok2->next()->insertToken(";");
tok2->next()->insertToken(value);
tok2 = tok2->tokAt(4);
}
}
}
}
// Binary operators simplification map
static const std::map<std::string, std::string> cAlternativeTokens = {
std::make_pair("and", "&&")
, std::make_pair("and_eq", "&=")
, std::make_pair("bitand", "&")
, std::make_pair("bitor", "|")
, std::make_pair("not_eq", "!=")
, std::make_pair("or", "||")
, std::make_pair("or_eq", "|=")
, std::make_pair("xor", "^")
, std::make_pair("xor_eq", "^=")
};
// Simplify the C alternative tokens:
// and => &&
// and_eq => &=
// bitand => &
// bitor => |
// compl => ~
// not => !
// not_eq => !=
// or => ||
// or_eq => |=
// xor => ^
// xor_eq => ^=
bool Tokenizer::simplifyCAlternativeTokens()
{
/* executable scope level */
unsigned int executableScopeLevel = 0;
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "{") {
if (executableScopeLevel > 0 || Token::simpleMatch(tok->previous(), ") {"))
++executableScopeLevel;
continue;
}
if (tok->str() == "}") {
if (executableScopeLevel > 0)
--executableScopeLevel;
continue;
}
if (!tok->isName())
continue;
const std::map<std::string, std::string>::const_iterator cOpIt = cAlternativeTokens.find(tok->str());
if (cOpIt != cAlternativeTokens.end()) {
if (!Token::Match(tok->previous(), "%name%|%num%|%char%|)|]|> %name% %name%|%num%|%char%|%op%|("))
continue;
if (Token::Match(tok->next(), "%assign%|%or%|%oror%|&&|*|/|%|^"))
continue;
tok->str(cOpIt->second);
ret = true;
} else if (Token::Match(tok, "not|compl")) {
// Don't simplify 'not p;' (in case 'not' is a type)
if (!Token::Match(tok->next(), "%name%|(") ||
Token::Match(tok->previous(), "[;{}]") ||
(executableScopeLevel == 0U && tok->strAt(-1) == "("))
continue;
tok->str((tok->str() == "not") ? "!" : "~");
ret = true;
}
}
return ret;
}
// int i(0); => int i; i = 0;
// int i(0), j; => int i; i = 0; int j;
void Tokenizer::simplifyInitVar()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName() || (tok->previous() && !Token::Match(tok->previous(), "[;{}]")))
continue;
if (tok->str() == "return")
continue;
if (Token::Match(tok, "class|struct|union| %type% *| %name% ( &| %any% ) ;")) {
tok = initVar(tok);
} else if (Token::Match(tok, "%type% *| %name% ( %type% (")) {
const Token* tok2 = tok->tokAt(2);
if (!tok2->link())
tok2 = tok2->next();
if (!tok2->link() || (tok2->link()->strAt(1) == ";" && !Token::simpleMatch(tok2->linkAt(2), ") (")))
tok = initVar(tok);
} else if (Token::Match(tok, "class|struct|union| %type% *| %name% ( &| %any% ) ,")) {
Token *tok1 = tok->tokAt(5);
while (tok1->str() != ",")
tok1 = tok1->next();
tok1->str(";");
const unsigned int numTokens = (Token::Match(tok, "class|struct|union")) ? 2U : 1U;
list.insertTokens(tok1, tok, numTokens);
tok = initVar(tok);
}
}
}
Token * Tokenizer::initVar(Token * tok)
{
// call constructor of class => no simplification
if (Token::Match(tok, "class|struct|union")) {
if (tok->strAt(2) != "*")
return tok;
tok = tok->next();
} else if (!tok->isStandardType() && tok->next()->str() != "*")
return tok;
// goto variable name..
tok = tok->next();
if (tok->str() == "*")
tok = tok->next();
// sizeof is not a variable name..
if (tok->str() == "sizeof")
return tok;
// check initializer..
if (tok->tokAt(2)->isStandardType() || tok->strAt(2) == "void")
return tok;
else if (!tok->tokAt(2)->isNumber() && !Token::Match(tok->tokAt(2), "%type% (") && tok->strAt(2) != "&" && tok->tokAt(2)->varId() == 0)
return tok;
// insert '; var ='
tok->insertToken(";");
tok->next()->insertToken(tok->str());
tok->tokAt(2)->varId(tok->varId());
tok = tok->tokAt(2);
tok->insertToken("=");
// goto '('..
tok = tok->tokAt(2);
// delete ')'
tok->link()->deleteThis();
// delete this
tok->deleteThis();
return tok;
}
bool Tokenizer::simplifyKnownVariables()
{
// return value for function. Set to true if any simplifications are made
bool ret = false;
// constants..
{
std::unordered_map<unsigned int, std::string> constantValues;
std::map<unsigned int, Token*> constantVars;
std::unordered_map<unsigned int, std::list<Token*>> constantValueUsages;
bool goback = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (goback) {
tok = tok->previous();
goback = false;
}
// Reference to variable
if (Token::Match(tok, "%type%|* & %name% = %name% ;")) {
Token *start = tok->previous();
while (Token::Match(start,"%type%|*|&"))
start = start->previous();
if (!Token::Match(start,"[;{}]"))
continue;
const Token *reftok = tok->tokAt(2);
const Token *vartok = reftok->tokAt(2);
int level = 0;
for (Token *tok2 = tok->tokAt(6); tok2; tok2 = tok2->next()) {
if (tok2->str() == "{") {
++level;
} else if (tok2->str() == "}") {
if (level <= 0)
break;
--level;
} else if (tok2->varId() == reftok->varId()) {
tok2->str(vartok->str());
tok2->varId(vartok->varId());
}
}
Token::eraseTokens(start, tok->tokAt(6));
tok = start;
}
if (tok->isName() && (Token::Match(tok, "static| const| static| %type% const| %name% = %any% ;") ||
Token::Match(tok, "static| const| static| %type% const| %name% ( %any% ) ;"))) {
bool isconst = false;
for (const Token *tok2 = tok; (tok2->str() != "=") && (tok2->str() != "("); tok2 = tok2->next()) {
if (tok2->str() == "const") {
isconst = true;
break;
}
}
if (!isconst)
continue;
Token *tok1 = tok;
// start of statement
if (tok != list.front() && !Token::Match(tok->previous(),";|{|}|private:|protected:|public:"))
continue;
// skip "const" and "static"
while (Token::Match(tok, "const|static"))
tok = tok->next();
// pod type
if (!tok->isStandardType())
continue;
Token * const vartok = (tok->next() && tok->next()->str() == "const") ? tok->tokAt(2) : tok->next();
const Token * const valuetok = vartok->tokAt(2);
if (Token::Match(valuetok, "%bool%|%char%|%num%|%str% )| ;")) {
// record a constant value for this variable
constantValues[vartok->varId()] = valuetok->str();
constantVars[vartok->varId()] = tok1;
}
} else if (tok->varId()) {
// find the entry for the known variable, if any. Exclude the location where the variable is assigned with next == "="
if (constantValues.find(tok->varId()) != constantValues.end() && tok->next()->str() != "=") {
constantValueUsages[tok->varId()].push_back(tok);
}
}
}
for (auto constantVar = constantVars.rbegin(); constantVar != constantVars.rend(); constantVar++) {
bool referenceFound = false;
std::list<Token*> usageList = constantValueUsages[constantVar->first];
for (Token* usage : usageList) {
// check if any usages of each known variable are a reference
if (Token::Match(usage->tokAt(-2), "(|[|,|{|return|%op% & %varid%", constantVar->first)) {
referenceFound = true;
break;
}
}
if (!referenceFound) {
// replace all usages of non-referenced known variables with their value
for (Token* usage : usageList) {
usage->str(constantValues[constantVar->first]);
}
Token* startTok = constantVar->second;
// remove variable assignment statement
while (startTok->next()->str() != ";")
startTok->deleteNext();
startTok->deleteNext();
// #8579 if we can we want another token to delete startTok. if we can't it doesn't matter
if (startTok->previous()) {
startTok->previous()->deleteNext();
} else if (startTok->next()) {
startTok->next()->deletePrevious();
} else {
startTok->deleteThis();
}
startTok = nullptr;
constantVar->second = nullptr;
ret = true;
}
}
}
// variable id for local, float/double, array variables
std::set<unsigned int> localvars;
std::set<unsigned int> floatvars;
std::set<unsigned int> arrays;
// auto variables..
for (Token *tok = list.front(); tok; tok = tok->next()) {
// Search for a block of code
Token * const start = const_cast<Token *>(startOfExecutableScope(tok));
if (!start)
continue;
for (const Token *tok2 = start->previous(); tok2 && !Token::Match(tok2, "[;{}]"); tok2 = tok2->previous()) {
if (tok2->varId() != 0)
localvars.insert(tok2->varId());
}
tok = start;
// parse the block of code..
int indentlevel = 0;
Token *tok2 = tok;
for (; tok2; tok2 = tok2->next()) {
if (Token::Match(tok2, "[;{}] %type% %name%|*")) {
bool isfloat = false;
bool ispointer = false;
const Token *vartok = tok2->next();
while (Token::Match(vartok, "%name%|* %name%|*")) {
if (Token::Match(vartok, "float|double"))
isfloat = true;
if (vartok->str() == "*")
ispointer = true;
vartok = vartok->next();
}
if (Token::Match(vartok, "%var% ;|["))
localvars.insert(vartok->varId());
if (isfloat && !ispointer && Token::Match(vartok, "%var% ;"))
floatvars.insert(vartok->varId());
if (Token::Match(vartok, "%var% ["))
arrays.insert(vartok->varId());
}
if (tok2->str() == "{")
++indentlevel;
else if (tok2->str() == "}") {
--indentlevel;
if (indentlevel <= 0)
break;
}
else if (Token::simpleMatch(tok2, "for ("))
tok2 = tok2->next()->link();
else if (tok2->previous()->str() != "*" && !Token::Match(tok2->tokAt(-2), "* --|++") &&
(Token::Match(tok2, "%name% = %bool%|%char%|%num%|%str%|%name% ;") ||
Token::Match(tok2, "%name% [ %num%| ] = %str% ;") ||
Token::Match(tok2, "%name% = & %name% ;") ||
(Token::Match(tok2, "%name% = & %name% [ 0 ] ;") && arrays.find(tok2->tokAt(3)->varId()) != arrays.end()))) {
const unsigned int varid = tok2->varId();
if (varid == 0)
continue;
if (Token::Match(tok2->previous(), "[;{}]") && localvars.find(varid) == localvars.end())
continue;
// initialization of static variable => the value is not *known*
{
bool isstatic = false;
const Token *decl = tok2->previous();
while (decl && (decl->isName() || decl->str() == "*")) {
if (decl->str() == "static") {
isstatic = true;
break;
}
decl = decl->previous();
}
if (isstatic)
continue;
}
// skip loop variable
if (Token::Match(tok2->tokAt(-2), "(|:: %type%")) {
const Token *tok3 = tok2->previous();
do {
tok3 = tok3->tokAt(-2);
} while (Token::Match(tok3->previous(), ":: %type%"));
if (Token::Match(tok3->tokAt(-2), "for ( %type%"))
continue;
}
// struct name..
if (Token::Match(tok2, "%varid% = &| %varid%", tok2->varId()))
continue;
const std::string structname = Token::Match(tok2->tokAt(-3), "[;{}] %name% .") ?
std::string(tok2->strAt(-2) + " .") :
std::string();
const Token * const valueToken = tok2->tokAt(2);
std::string value;
unsigned int valueVarId = 0;
Token *tok3 = nullptr;
bool valueIsPointer = false;
// there could be a hang here if tok2 is moved back by the function calls below for some reason
if (mSettings->terminated())
return false;
if (!simplifyKnownVariablesGetData(varid, &tok2, &tok3, value, valueVarId, valueIsPointer, floatvars.find(tok2->varId()) != floatvars.end()))
continue;
if (valueVarId > 0 && arrays.find(valueVarId) != arrays.end())
continue;
ret |= simplifyKnownVariablesSimplify(&tok2, tok3, varid, structname, value, valueVarId, valueIsPointer, valueToken, indentlevel);
}
else if (Token::Match(tok2, "strcpy|sprintf ( %name% , %str% ) ;")) {
const unsigned int varid(tok2->tokAt(2)->varId());
if (varid == 0)
continue;
const Token * const valueToken = tok2->tokAt(4);
std::string value(valueToken->str());
if (tok2->str() == "sprintf") {
std::string::size_type n = 0;
while ((n = value.find("%%", n)) != std::string::npos) {
// Replace "%%" with "%" - erase the first '%' and continue past the second '%'
value.erase(n, 1);
++n;
}
}
const unsigned int valueVarId(0);
const bool valueIsPointer(false);
Token *tok3 = tok2->tokAt(6);
ret |= simplifyKnownVariablesSimplify(&tok2, tok3, varid, emptyString, value, valueVarId, valueIsPointer, valueToken, indentlevel);
// there could be a hang here if tok2 was moved back by the function call above for some reason
if (mSettings->terminated())
return false;
}
}
if (tok2)
tok = tok2->previous();
}
return ret;
}
bool Tokenizer::simplifyKnownVariablesGetData(unsigned int varid, Token **_tok2, Token **_tok3, std::string &value, unsigned int &valueVarId, bool &valueIsPointer, bool floatvar)
{
Token *tok2 = *_tok2;
Token *tok3 = nullptr;
if (Token::simpleMatch(tok2->tokAt(-2), "for (")) {
// only specific for loops is handled
if (!Token::Match(tok2, "%varid% = %num% ; %varid% <|<= %num% ; ++| %varid% ++| ) {", varid))
return false;
// is there a "break" in the for loop?
bool hasbreak = false;
const Token* end4 = tok2->linkAt(-1)->linkAt(1);
for (const Token *tok4 = tok2->previous()->link(); tok4 != end4; tok4 = tok4->next()) {
if (tok4->str() == "break") {
hasbreak = true;
break;
}
}
if (hasbreak)
return false;
// no break => the value of the counter value is known after the for loop..
const Token* compareTok = tok2->tokAt(5);
if (compareTok->str() == "<") {
value = compareTok->next()->str();
valueVarId = compareTok->next()->varId();
} else
value = MathLib::toString(MathLib::toLongNumber(compareTok->next()->str()) + 1);
// Skip for-body..
tok3 = tok2->previous()->link()->next()->link()->next();
} else {
value = tok2->strAt(2);
valueVarId = tok2->tokAt(2)->varId();
if (tok2->strAt(1) == "[") {
value = tok2->next()->link()->strAt(2);
valueVarId = 0;
} else if (value == "&") {
value = tok2->strAt(3);
valueVarId = tok2->tokAt(3)->varId();
// *ptr = &var; *ptr = 5;
// equals
// var = 5; not *var = 5;
if (tok2->strAt(4) == ";")
valueIsPointer = true;
}
// Add a '.0' to a decimal value and therefore convert it to an floating point number.
else if (MathLib::isDec(tok2->strAt(2)) && floatvar) {
value += ".0";
}
// float variable: convert true/false to 1.0 / 0.0
else if (tok2->tokAt(2)->isBoolean() && floatvar) {
value = (value == "true") ? "1.0" : "0.0";
}
if (Token::simpleMatch(tok2->next(), "= &"))
tok2 = tok2->tokAt(3);
tok3 = tok2->next();
}
*_tok2 = tok2;
*_tok3 = tok3;
return true;
}
bool Tokenizer::simplifyKnownVariablesSimplify(Token **tok2, Token *tok3, unsigned int varid, const std::string &structname, std::string &value, unsigned int valueVarId, bool valueIsPointer, const Token * const valueToken, int indentlevel) const
{
const bool pointeralias(valueToken->isName() || Token::Match(valueToken, "& %name% ["));
const bool varIsGlobal = (indentlevel == 0);
const bool printDebug = mSettings->debugwarnings;
if (mErrorLogger && !list.getFiles().empty())
mErrorLogger->reportProgress(list.getFiles()[0], "Tokenize (simplifyKnownVariables)", tok3->progressValue());
if (isMaxTime())
return false;
bool ret = false;
Token* bailOutFromLoop = nullptr;
int indentlevel3 = indentlevel;
bool ret3 = false;
for (; tok3; tok3 = tok3->next()) {
if (tok3->str() == "{") {
++indentlevel3;
} else if (tok3->str() == "}") {
--indentlevel3;
if (indentlevel3 < indentlevel) {
if (Token::Match((*tok2)->tokAt(-7), "%type% * %name% ; %name% = & %name% ;") &&
(*tok2)->strAt(-5) == (*tok2)->strAt(-3)) {
(*tok2) = (*tok2)->tokAt(-4);
Token::eraseTokens((*tok2), (*tok2)->tokAt(6));
}
break;
}
}
// Stop if there is a pointer alias and a shadow variable is
// declared in an inner scope (#3058)
if (valueIsPointer && tok3->varId() > 0 &&
tok3->previous() && (tok3->previous()->isName() || tok3->previous()->str() == "*") &&
valueToken->str() == "&" &&
valueToken->next() &&
valueToken->next()->isName() &&
tok3->str() == valueToken->next()->str() &&
tok3->varId() > valueToken->next()->varId()) {
// more checking if this is a variable declaration
bool decl = true;
for (const Token *tok4 = tok3->previous(); tok4; tok4 = tok4->previous()) {
if (Token::Match(tok4, "[;{}]"))
break;
else if (tok4->isName()) {
if (tok4->varId() > 0) {
decl = false;
break;
}
}
else if (!Token::Match(tok4, "[&*]")) {
decl = false;
break;
}
}
if (decl)
break;
}
// Stop if label is found
if (Token::Match(tok3, "; %type% : ;"))
break;
// Stop if break/continue is found ..
if (Token::Match(tok3, "break|continue"))
break;
if ((indentlevel3 > 1 || !Token::simpleMatch(Token::findsimplematch(tok3,";"), "; }")) && tok3->str() == "return")
ret3 = true;
if (ret3 && tok3->str() == ";")
break;
if (pointeralias && Token::Match(tok3, ("!!= " + value).c_str()))
break;
// Stop if a loop is found
if (pointeralias && Token::Match(tok3, "do|for|while"))
break;
// Stop if unknown function call is seen and the variable is global: it might be
// changed by the function call
if (varIsGlobal && tok3->str() == ")" && tok3->link() &&
Token::Match(tok3->link()->tokAt(-2), "[;{}] %name% (") &&
!Token::Match(tok3->link()->previous(), "if|for|while|switch|BOOST_FOREACH"))
break;
// Stop if something like 'while (--var)' is found
if (Token::Match(tok3, "for|while|do")) {
const Token *endpar = tok3->next()->link();
if (Token::simpleMatch(endpar, ") {"))
endpar = endpar->next()->link();
bool bailout = false;
for (const Token *tok4 = tok3; tok4 && tok4 != endpar; tok4 = tok4->next()) {
if (Token::Match(tok4, "++|-- %varid%", varid) ||
Token::Match(tok4, "%varid% ++|--|=", varid)) {
bailout = true;
break;
}
}
if (bailout)
break;
}
if (bailOutFromLoop) {
// This could be a loop, skip it, but only if it doesn't contain
// the variable we are checking for. If it contains the variable
// we will bail out.
if (tok3->varId() == varid) {
// Continue
//tok2 = bailOutFromLoop;
break;
} else if (tok3 == bailOutFromLoop) {
// We have skipped the loop
bailOutFromLoop = nullptr;
continue;
}
continue;
} else if (tok3->str() == "{" && tok3->previous()->str() == ")") {
// There is a possible loop after the assignment. Try to skip it.
if (tok3->previous()->link() &&
tok3->previous()->link()->strAt(-1) != "if")
bailOutFromLoop = tok3->link();
continue;
}
// Variable used in realloc (see Ticket #1649)
if (Token::Match(tok3, "%name% = realloc ( %name% ,") &&
tok3->varId() == varid &&
tok3->tokAt(4)->varId() == varid) {
tok3->tokAt(4)->str(value);
ret = true;
}
// condition "(|&&|%OROR% %varid% )|&&|%OROR%|;
if (!Token::Match(tok3->previous(), "( %name% )") &&
Token::Match(tok3->previous(), "&&|(|%oror% %varid% &&|%oror%|)|;", varid)) {
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
// parameter in function call..
if (tok3->varId() == varid && Token::Match(tok3->previous(), "[(,] %name% [,)]")) {
// If the parameter is passed by value then simplify it
if (isFunctionParameterPassedByValue(tok3)) {
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
}
// Variable is used somehow in a non-defined pattern => bail out
if (tok3->varId() == varid) {
// This is a really generic bailout so let's try to avoid this.
// There might be lots of false negatives.
if (printDebug) {
// FIXME: Fix all the debug warnings for values and then
// remove this bailout
if (pointeralias)
break;
// suppress debug-warning when calling member function
if (Token::Match(tok3->next(), ". %name% ("))
break;
// suppress debug-warning when assignment
if (tok3->strAt(1) == "=")
break;
// taking address of variable..
if (Token::Match(tok3->tokAt(-2), "return|= & %name% ;"))
break;
// parameter in function call..
if (Token::Match(tok3->tokAt(-2), "%name% ( %name% ,|)") ||
Token::Match(tok3->previous(), ", %name% ,|)"))
break;
// conditional increment
if (Token::Match(tok3->tokAt(-3), ") { ++|--") ||
Token::Match(tok3->tokAt(-2), ") { %name% ++|--"))
break;
reportError(tok3, Severity::debug, "debug",
"simplifyKnownVariables: bailing out (variable="+tok3->str()+", value="+value+")");
}
break;
}
// Using the variable in condition..
if (Token::Match(tok3->previous(), ("if ( " + structname + " %varid% %cop%|)").c_str(), varid) ||
Token::Match(tok3, ("( " + structname + " %varid% %comp%").c_str(), varid) ||
Token::Match(tok3, ("%comp%|!|= " + structname + " %varid% %cop%|)|;").c_str(), varid) ||
Token::Match(tok3->previous(), "strlen|free ( %varid% )", varid)) {
if (value[0] == '\"' && tok3->previous()->str() != "strlen") {
// bail out if value is a string unless if it's just given
// as parameter to strlen
break;
}
if (!structname.empty()) {
tok3->deleteNext(2);
}
if (Token::Match(valueToken, "& %name% ;")) {
tok3->insertToken("&");
tok3 = tok3->next();
}
tok3 = tok3->next();
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
// pointer alias used in condition..
if (Token::Match(valueToken,"& %name% ;") && Token::Match(tok3, ("( * " + structname + " %varid% %cop%").c_str(), varid)) {
tok3->deleteNext();
if (!structname.empty())
tok3->deleteNext(2);
tok3 = tok3->next();
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
// Delete pointer alias
if (isCPP() && pointeralias && (tok3->str() == "delete") && tok3->next() &&
(Token::Match(tok3->next(), "%varid% ;", varid) ||
Token::Match(tok3->next(), "[ ] %varid%", varid))) {
tok3 = (tok3->next()->str() == "[") ? tok3->tokAt(3) : tok3->next();
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
// Variable is used in function call..
if (Token::Match(tok3, ("%name% ( " + structname + " %varid% ,").c_str(), varid)) {
static const char * const functionName[] = {
// always simplify
"strcmp", "strdup",
// don't simplify buffer value
"memcmp","memcpy","memmove","memset","strcpy","strncmp","strncpy"
};
for (unsigned int i = 0; i < (sizeof(functionName) / sizeof(*functionName)); ++i) {
if (valueVarId == 0U && i >= 2)
break;
if (tok3->str() == functionName[i]) {
Token *par1 = tok3->tokAt(2);
if (!structname.empty()) {
par1->deleteNext();
par1->deleteThis();
}
par1->str(value);
par1->varId(valueVarId);
break;
}
}
}
// Variable is used as 2nd parameter in function call..
if (Token::Match(tok3, ("%name% ( %any% , " + structname + " %varid% ,|)").c_str(), varid)) {
static const char * const functionName[] = {
// always simplify
"strcmp","strcpy","strncmp","strncpy",
// don't simplify buffer value
"memcmp","memcpy","memmove"
};
for (unsigned int i = 0; i < (sizeof(functionName) / sizeof(*functionName)); ++i) {
if (valueVarId == 0U && i >= 4)
break;
if (tok3->str() == functionName[i]) {
Token *par = tok3->tokAt(4);
if (!structname.empty()) {
par->deleteNext();
par->deleteThis();
}
par->str(value);
par->varId(valueVarId);
break;
}
}
}
// array usage
if (value[0] != '\"' && Token::Match(tok3, ("[(,] " + structname + " %varid% [|%cop%").c_str(), varid)) {
if (!structname.empty()) {
tok3->deleteNext(2);
}
tok3 = tok3->next();
tok3->str(value);
tok3->varId(valueVarId);
ret = true;
}
// The >> operator is sometimes used to assign a variable in C++
if (isCPP() && Token::Match(tok3, (">> " + structname + " %varid%").c_str(), varid)) {
// bailout for such code: ; std :: cin >> i ;
const Token *prev = tok3->previous();
while (prev && prev->str() != "return" && Token::Match(prev, "%name%|::|*"))
prev = prev->previous();
if (Token::Match(prev, ";|{|}|>>"))
break;
}
// Variable is used in calculation..
if (((tok3->previous()->varId() > 0) && Token::Match(tok3, ("& " + structname + " %varid%").c_str(), varid)) ||
(Token::Match(tok3, ("[=+-*/%^|[] " + structname + " %varid% [=?+-*/%^|;])]").c_str(), varid) && !Token::Match(tok3, ("= " + structname + " %name% =").c_str())) ||
Token::Match(tok3, ("[(=+-*/%^|[] " + structname + " %varid% <<|>>").c_str(), varid) ||
Token::Match(tok3, ("<<|>> " + structname + " %varid% %cop%|;|]|)").c_str(), varid) ||
Token::Match(tok3->previous(), ("[=+-*/%^|[] ( " + structname + " %varid% !!=").c_str(), varid)) {
if (value[0] == '\"')
break;
if (!structname.empty()) {
tok3->deleteNext(2);
ret = true;
}
tok3 = tok3->next();
if (tok3->str() != value)
ret = true;
tok3->str(value);
tok3->varId(valueVarId);
if (tok3->previous()->str() == "*" && (valueIsPointer || Token::Match(valueToken, "& %name% ;"))) {
tok3 = tok3->previous();
tok3->deleteThis();
ret = true;
} else if (Token::Match(valueToken, "& %name% ;"))
tok3->insertToken("&", emptyString, true);
}
if (Token::simpleMatch(tok3, "= {")) {
const Token* const end4 = tok3->linkAt(1);
for (const Token *tok4 = tok3; tok4 != end4; tok4 = tok4->next()) {
if (Token::Match(tok4, "{|, %varid% ,|}", varid)) {
tok4->next()->str(value);
tok4->next()->varId(valueVarId);
ret = true;
}
}
}
// Using the variable in for-condition..
if (Token::simpleMatch(tok3, "for (")) {
for (Token *tok4 = tok3->tokAt(2); tok4; tok4 = tok4->next()) {
if (Token::Match(tok4, "(|)"))
break;
// Replace variable used in condition..
if (Token::Match(tok4, "; %name% <|<=|!= %name% ; ++| %name% ++| )")) {
const Token *inctok = tok4->tokAt(5);
if (inctok->str() == "++")
inctok = inctok->next();
if (inctok->varId() == varid)
break;
if (tok4->next()->varId() == varid) {
tok4->next()->str(value);
tok4->next()->varId(valueVarId);
ret = true;
}
if (tok4->tokAt(3)->varId() == varid) {
tok4->tokAt(3)->str(value);
tok4->tokAt(3)->varId(valueVarId);
ret = true;
}
}
}
}
if (indentlevel == indentlevel3 && Token::Match(tok3->next(), "%varid% ++|--", varid) && MathLib::isInt(value)) {
const std::string op(tok3->strAt(2));
if (Token::Match(tok3, "[{};] %any% %any% ;")) {
tok3->deleteNext(3);
} else {
tok3 = tok3->next();
tok3->str(value);
tok3->varId(valueVarId);
tok3->deleteNext();
}
value = MathLib::incdec(value, op);
if (!Token::simpleMatch((*tok2)->tokAt(-2), "for (")) {
(*tok2)->tokAt(2)->str(value);
(*tok2)->tokAt(2)->varId(valueVarId);
}
ret = true;
}
if (indentlevel == indentlevel3 && Token::Match(tok3->next(), "++|-- %varid%", varid) && MathLib::isInt(value) &&
!Token::Match(tok3->tokAt(3), "[.[]")) {
value = MathLib::incdec(value, tok3->next()->str());
(*tok2)->tokAt(2)->str(value);
(*tok2)->tokAt(2)->varId(valueVarId);
if (Token::Match(tok3, "[;{}] %any% %any% ;")) {
tok3->deleteNext(3);
} else {
tok3->deleteNext();
tok3->next()->str(value);
tok3->next()->varId(valueVarId);
}
tok3 = tok3->next();
ret = true;
}
// return variable..
if (Token::Match(tok3, "return %varid% %any%", varid) &&
valueToken->str() != "&" &&
(tok3->tokAt(2)->isExtendedOp() || tok3->strAt(2) == ";") &&
value[0] != '\"') {
tok3->next()->str(value);
tok3->next()->varId(valueVarId);
}
else if (pointeralias && Token::Match(tok3, "return * %varid% ;", varid) && value[0] != '\"') {
tok3->deleteNext();
tok3->next()->str(value);
tok3->next()->varId(valueVarId);
}
}
return ret;
}
void Tokenizer::elseif()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::simpleMatch(tok, "else if"))
continue;
for (Token *tok2 = tok; tok2; tok2 = tok2->next()) {
if (Token::Match(tok2, "(|{|["))
tok2 = tok2->link();
if (Token::Match(tok2, "}|;")) {
if (tok2->next() && tok2->next()->str() != "else") {
tok->insertToken("{");
tok2->insertToken("}");
Token::createMutualLinks(tok->next(), tok2->next());
break;
}
}
}
}
}
bool Tokenizer::simplifyRedundantParentheses()
{
bool ret = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() != "(")
continue;
if (Token::simpleMatch(tok, "( {"))
continue;
if (Token::Match(tok->link(), ") %num%")) {
tok = tok->link();
continue;
}
// !!operator = ( x ) ;
if (tok->strAt(-2) != "operator" &&
tok->previous() && tok->previous()->str() == "=" &&
tok->next() && tok->next()->str() != "{" &&
Token::simpleMatch(tok->link(), ") ;")) {
tok->link()->deleteThis();
tok->deleteThis();
continue;
}
while (Token::simpleMatch(tok, "( (") &&
tok->link() && tok->link()->previous() == tok->next()->link()) {
// We have "(( *something* ))", remove the inner
// parentheses
tok->deleteNext();
tok->link()->tokAt(-2)->deleteNext();
ret = true;
}
if (isCPP() && Token::Match(tok->tokAt(-2), "[;{}=(] new (") && Token::Match(tok->link(), ") [;,{}[]")) {
// Remove the parentheses in "new (type)" constructs
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (Token::Match(tok->previous(), "! ( %name% )")) {
// Remove the parentheses
tok->deleteThis();
tok->deleteNext();
ret = true;
}
if (Token::Match(tok->previous(), "[(,;{}] ( %name% ) .")) {
// Remove the parentheses
tok->deleteThis();
tok->deleteNext();
ret = true;
}
if (Token::Match(tok->previous(), "[(,;{}] ( %name% (") &&
tok->link()->previous() == tok->linkAt(2)) {
// We have "( func ( *something* ))", remove the outer
// parentheses
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (Token::Match(tok->previous(), "[,;{}] ( delete [| ]| %name% ) ;")) {
// We have "( delete [| ]| var )", remove the outer
// parentheses
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (!Token::simpleMatch(tok->tokAt(-2), "operator delete") &&
Token::Match(tok->previous(), "delete|; (") &&
(tok->previous()->str() != "delete" || tok->next()->varId() > 0) &&
Token::Match(tok->link(), ") ;|,")) {
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (Token::Match(tok->previous(), "[(!*;{}] ( %name% )") &&
(tok->next()->varId() != 0 || Token::Match(tok->tokAt(3), "[+-/=]")) && !tok->next()->isStandardType()) {
// We have "( var )", remove the parentheses
tok->deleteThis();
tok->deleteNext();
ret = true;
}
while (Token::Match(tok->previous(), "[;{}[(,!*] ( %name% .")) {
Token *tok2 = tok->tokAt(2);
while (Token::Match(tok2, ". %name%")) {
tok2 = tok2->tokAt(2);
}
if (tok2 != tok->link())
break;
// We have "( var . var . ... . var )", remove the parentheses
tok = tok->previous();
tok->deleteNext();
tok2->deleteThis();
ret = true;
continue;
}
if (Token::simpleMatch(tok->previous(), "? (") && Token::simpleMatch(tok->link(), ") :")) {
const Token *tok2 = tok->next();
while (tok2 && (Token::Match(tok2,"%bool%|%num%|%name%") || tok2->isArithmeticalOp()))
tok2 = tok2->next();
if (tok2 && tok2->str() == ")") {
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
continue;
}
}
while (Token::Match(tok->previous(), "[{([,] ( !!{") &&
Token::Match(tok->link(), ") [;,])]") &&
!Token::simpleMatch(tok->tokAt(-2), "operator ,") && // Ticket #5709
!Token::findsimplematch(tok, ",", tok->link())) {
// We have "( ... )", remove the parentheses
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (Token::simpleMatch(tok->previous(), ", (") &&
Token::simpleMatch(tok->link(), ") =")) {
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
// Simplify "!!operator !!%name%|)|>|>> ( %num%|%bool% ) %op%|;|,|)"
if (Token::Match(tok, "( %bool%|%num% ) %cop%|;|,|)") &&
tok->strAt(-2) != "operator" &&
tok->previous() &&
!Token::Match(tok->previous(), "%name%|)") &&
(!(isCPP() && Token::Match(tok->previous(),">|>>")))) {
tok->link()->deleteThis();
tok->deleteThis();
ret = true;
}
if (Token::Match(tok->previous(), "*|& ( %name% )")) {
// We may have a variable declaration looking like "type_name *(var_name)"
Token *tok2 = tok->tokAt(-2);
while (Token::Match(tok2, "%type%|static|const|extern") && tok2->str() != "operator") {
tok2 = tok2->previous();
}
if (tok2 && !Token::Match(tok2, "[;,{]")) {
// Not a variable declaration
} else {
tok->deleteThis();
tok->deleteNext();
}
}
}
return ret;
}
void Tokenizer::simplifyCharAt()
{
// Replace "string"[0] with 's'
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%str% [ %num% ]")) {
const MathLib::bigint index = MathLib::toLongNumber(tok->strAt(2));
// Check within range
if (index >= 0 && index <= Token::getStrLength(tok)) {
tok->str("'" + Token::getCharAt(tok, index) + "'");
tok->deleteNext(3);
}
}
}
}
void Tokenizer::simplifyReference()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
// starting executable scope..
Token *start = const_cast<Token *>(startOfExecutableScope(tok));
if (start) {
tok = start;
// replace references in this scope..
Token * const end = tok->link();
for (Token *tok2 = tok; tok2 && tok2 != end; tok2 = tok2->next()) {
// found a reference..
if (Token::Match(tok2, "[;{}] %type% & %name% (|= %name% )| ;")) {
const unsigned int refId = tok2->tokAt(3)->varId();
if (!refId)
continue;
// replace reference in the code..
for (Token *tok3 = tok2->tokAt(7); tok3 && tok3 != end; tok3 = tok3->next()) {
if (tok3->varId() == refId) {
tok3->str(tok2->strAt(5));
tok3->varId(tok2->tokAt(5)->varId());
}
}
tok2->deleteNext(6+(tok2->strAt(6)==")" ? 1 : 0));
}
}
tok = end;
}
}
}
bool Tokenizer::simplifyCalculations()
{
return mTemplateSimplifier->simplifyCalculations();
}
void Tokenizer::simplifyOffsetPointerDereference()
{
// Replace "*(str + num)" => "str[num]" and
// Replace "*(str - num)" => "str[-num]"
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName() && !tok->isLiteral()
&& !Token::Match(tok, "]|)|++|--")
&& Token::Match(tok->next(), "* ( %name% +|- %num%|%name% )")) {
// remove '* ('
tok->deleteNext(2);
// '+'->'['
tok = tok->tokAt(2);
Token* const openBraceTok = tok;
const bool isNegativeIndex = (tok->str() == "-");
tok->str("[");
// Insert a "-" in front of the number or variable
if (isNegativeIndex) {
if (tok->next()->isName()) {
tok->insertToken("-");
tok = tok->next();
} else
tok->next()->str(std::string("-") + tok->next()->str());
}
tok = tok->tokAt(2);
tok->str("]");
Token::createMutualLinks(openBraceTok, tok);
}
}
}
void Tokenizer::simplifyOffsetPointerReference()
{
std::set<unsigned int> pod;
for (const Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->isStandardType()) {
tok = tok->next();
while (tok && (tok->str() == "*" || tok->isName())) {
if (tok->varId() > 0) {
pod.insert(tok->varId());
break;
}
tok = tok->next();
}
if (!tok)
break;
}
}
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "%num%|%name%|]|)") &&
(Token::Match(tok->next(), "& %name% [ %num%|%name% ] !!["))) {
tok = tok->next();
if (tok->next()->varId()) {
if (pod.find(tok->next()->varId()) == pod.end()) {
tok = tok->tokAt(5);
if (!tok)
syntaxError(tok);
continue;
}
}
// '&' => '('
tok->str("(");
tok = tok->next();
// '[' => '+'
tok->deleteNext();
tok->insertToken("+");
tok = tok->tokAt(3);
//remove ']'
tok->str(")");
Token::createMutualLinks(tok->tokAt(-4), tok);
}
}
}
void Tokenizer::simplifyNestedStrcat()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (! Token::Match(tok, "[;{}] strcat ( strcat (")) {
continue;
}
// find inner strcat call
Token *tok2 = tok->tokAt(3);
while (Token::simpleMatch(tok2, "strcat ( strcat"))
tok2 = tok2->tokAt(2);
if (tok2->strAt(3) != ",")
continue;
// If we have this code:
// strcat(strcat(dst, foo), bar);
// We move this part of code before all strcat() calls: strcat(dst, foo)
// And place "dst" token where the code was.
Token *prevTok = tok2->previous();
// Move tokens to new place
Token::move(tok2, tok2->next()->link(), tok);
tok = tok2->next()->link();
// Insert the "dst" token
prevTok->insertToken(tok2->strAt(2));
prevTok->next()->varId(tok2->tokAt(2)->varId());
// Insert semicolon after the moved strcat()
tok->insertToken(";");
}
}
// Check if this statement is a duplicate definition. A duplicate
// definition will hide the enumerator within it's scope so just
// skip the entire scope of the duplicate.
bool Tokenizer::duplicateDefinition(Token ** tokPtr)
{
// check for an end of definition
const Token * tok = *tokPtr;
if (tok && Token::Match(tok->next(), ";|,|[|=|)|>")) {
const Token * end = tok->next();
if (end->str() == "[") {
end = end->link()->next();
} else if (end->str() == ",") {
// check for function argument
if (Token::Match(tok->previous(), "(|,"))
return false;
// find end of definition
int level = 0;
while (end->next() && (!Token::Match(end->next(), ";|)|>") ||
(end->next()->str() == ")" && level == 0))) {
if (end->next()->str() == "(")
++level;
else if (end->next()->str() == ")")
--level;
end = end->next();
}
} else if (end->str() == ")") {
// check for function argument
if (tok->previous()->str() == ",")
return false;
}
if (end) {
if (Token::simpleMatch(end, ") {")) { // function parameter ?
// make sure it's not a conditional
if (Token::Match(end->link()->previous(), "if|for|while|switch|BOOST_FOREACH") || Token::Match(end->link()->tokAt(-2), ":|,"))
return false;
// look backwards
if (tok->previous()->str() == "enum" ||
(Token::Match(tok->previous(), "%type%") &&
tok->previous()->str() != "return") ||
Token::Match(tok->tokAt(-2), "%type% &|*")) {
// duplicate definition so skip entire function
*tokPtr = end->next()->link();
return true;
}
} else if (end->str() == ">") { // template parameter ?
// look backwards
if (tok->previous()->str() == "enum" ||
(Token::Match(tok->previous(), "%type%") &&
tok->previous()->str() != "return")) {
// duplicate definition so skip entire template
while (end && end->str() != "{")
end = end->next();
if (end) {
*tokPtr = end->link();
return true;
}
}
} else {
if (Token::Match(tok->previous(), "enum|,"))
return true;
else if (Token::Match(tok->previous(), "%type%")) {
// look backwards
const Token *back = tok;
while (back && back->isName())
back = back->previous();
if (!back || (Token::Match(back, "[(,;{}]") && !Token::Match(back->next(),"return|throw")))
return true;
}
}
}
}
return false;
}
static const std::set<std::string> stdFunctionsPresentInC = {
"strcat",
"strcpy",
"strncat",
"strncpy",
"free",
"malloc",
"strdup"
};
void Tokenizer::simplifyStd()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() != "std")
continue;
if (Token::Match(tok->previous(), "[(,{};] std :: %name% (") &&
stdFunctionsPresentInC.find(tok->strAt(2)) != stdFunctionsPresentInC.end()) {
tok->deleteNext();
tok->deleteThis();
}
}
}
//---------------------------------------------------------------------------
// Helper functions for handling the tokens list
//---------------------------------------------------------------------------
//---------------------------------------------------------------------------
bool Tokenizer::IsScopeNoReturn(const Token *endScopeToken, bool *unknown) const
{
std::string unknownFunc;
const bool ret = mSettings->library.isScopeNoReturn(endScopeToken,&unknownFunc);
if (unknown)
*unknown = !unknownFunc.empty();
if (!unknownFunc.empty() && mSettings->checkLibrary && mSettings->isEnabled(Settings::INFORMATION)) {
// Is function global?
bool globalFunction = true;
if (Token::simpleMatch(endScopeToken->tokAt(-2), ") ; }")) {
const Token * const ftok = endScopeToken->linkAt(-2)->previous();
if (ftok &&
ftok->isName() &&
ftok->function() &&
ftok->function()->nestedIn &&
ftok->function()->nestedIn->type != Scope::eGlobal) {
globalFunction = false;
}
}
// don't warn for nonglobal functions (class methods, functions hidden in namespaces) since they can't be configured yet
// FIXME: when methods and namespaces can be configured properly, remove the "globalFunction" check
if (globalFunction) {
reportError(endScopeToken->previous(),
Severity::information,
"checkLibraryNoReturn",
"--check-library: Function " + unknownFunc + "() should have <noreturn> configuration");
}
}
return ret;
}
//---------------------------------------------------------------------------
bool Tokenizer::isFunctionParameterPassedByValue(const Token *fpar) const
{
// TODO: If symbol database is available, use it.
const Token *ftok;
// Look at function call, what parameter number is it?
unsigned int parameter = 1;
for (ftok = fpar->previous(); ftok; ftok = ftok->previous()) {
if (ftok->str() == "(")
break;
else if (ftok->str() == ")")
ftok = ftok->link();
else if (ftok->str() == ",")
++parameter;
else if (Token::Match(ftok, "[;{}]"))
break;
}
// Is this a function call?
if (ftok && Token::Match(ftok->tokAt(-2), "[;{}=] %name% (")) {
const std::string& functionName(ftok->previous()->str());
if (functionName == "return")
return true;
// Locate function declaration..
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (tok->str() == "{")
tok = tok->link();
else if (Token::Match(tok, "%type% (") && tok->str() == functionName) {
// Goto parameter
tok = tok->tokAt(2);
unsigned int par = 1;
while (tok && par < parameter) {
if (tok->str() == ")")
break;
if (tok->str() == ",")
++par;
tok = tok->next();
}
if (!tok)
return false;
// If parameter was found, determine if it's passed by value
if (par == parameter) {
bool knowntype = false;
while (tok && tok->isName()) {
knowntype |= tok->isStandardType();
knowntype |= (tok->str() == "struct");
tok = tok->next();
}
if (!tok || !knowntype)
return false;
if (tok->str() != "," && tok->str() != ")")
return false;
return true;
}
}
}
}
return false;
}
//---------------------------------------------------------------------------
void Tokenizer::eraseDeadCode(Token *begin, const Token *end)
{
if (!begin)
return;
const bool isgoto = Token::Match(begin->tokAt(-2), "goto %name% ;");
unsigned int indentlevel = 1,
indentcase = 0,
indentswitch = 0,
indentlabel = 0,
roundbraces = 0,
indentcheck = 0;
std::vector<unsigned int> switchindents;
bool checklabel = false;
Token *tok = begin;
Token *tokcheck = nullptr;
while (tok->next() && tok->next() != end) {
if (tok->next()->str() == "(") {
++roundbraces;
tok->deleteNext();
continue;
} else if (tok->next()->str() == ")") {
if (!roundbraces)
break; //too many ending round parentheses
--roundbraces;
tok->deleteNext();
continue;
}
if (roundbraces) {
tok->deleteNext();
continue;
}
if (Token::Match(tok, "[{};] switch (")) {
if (!checklabel) {
if (!indentlabel) {
//remove 'switch ( ... )'
Token::eraseTokens(tok, tok->linkAt(2)->next());
} else {
tok = tok->linkAt(2);
}
if (tok->next()->str() == "{") {
++indentswitch;
indentcase = indentlevel + 1;
switchindents.push_back(indentcase);
}
} else {
tok = tok->linkAt(2);
if (Token::simpleMatch(tok, ") {")) {
++indentswitch;
indentcase = indentlevel + 1;
switchindents.push_back(indentcase);
}
}
} else if (tok->next()->str() == "{") {
++indentlevel;
if (!checklabel) {
checklabel = true;
tokcheck = tok;
indentcheck = indentlevel;
indentlabel = 0;
}
tok = tok->next();
} else if (tok->next()->str() == "}") {
--indentlevel;
if (!indentlevel)
break;
if (!checklabel) {
tok->deleteNext();
} else {
if (indentswitch && indentlevel == indentcase)
--indentlevel;
if (indentlevel < indentcheck) {
const Token *end2 = tok->next();
tok = end2->link()->previous(); //return to initial '{'
if (indentswitch && Token::simpleMatch(tok, ") {") && Token::Match(tok->link()->tokAt(-2), "[{};] switch ("))
tok = tok->link()->tokAt(-2); //remove also 'switch ( ... )'
Token::eraseTokens(tok, end2->next());
checklabel = false;
tokcheck = nullptr;
indentcheck = 0;
} else {
tok = tok->next();
}
}
if (indentswitch && indentlevel <= indentcase) {
--indentswitch;
switchindents.pop_back();
if (!indentswitch)
indentcase = 0;
else
indentcase = switchindents[indentswitch-1];
}
} else if (Token::Match(tok, "[{};:] case")) {
const Token *tok2 = Token::findsimplematch(tok->next(), ": ;", end);
if (!tok2) {
tok->deleteNext();
continue;
}
if (indentlevel == 1)
break; //it seems like the function was called inside a case-default block.
if (indentlevel == indentcase)
++indentlevel;
tok2 = tok2->next();
if (!checklabel || !indentswitch) {
Token::eraseTokens(tok, tok2->next());
} else {
tok = const_cast<Token *>(tok2);
}
} else if (Token::Match(tok, "[{};] default : ;")) {
if (indentlevel == 1)
break; //it seems like the function was called inside a case-default block.
if (indentlevel == indentcase)
++indentlevel;
if (!checklabel || !indentswitch) {
tok->deleteNext(3);
} else {
tok = tok->tokAt(3);
}
} else if (Token::Match(tok, "[{};] %name% : ;") && tok->next()->str() != "default") {
if (checklabel) {
indentlabel = indentlevel;
tok = tokcheck->next();
checklabel = false;
indentlevel = indentcheck;
} else {
if (indentswitch) {
//Before stopping the function, since the 'switch()'
//instruction is removed, there's no sense to keep the
//case instructions. Remove them, if there are any.
Token *tok2 = tok->tokAt(3);
unsigned int indentlevel2 = indentlevel;
while (tok2->next() && tok2->next() != end) {
if (Token::Match(tok2->next(), "{|[|(")) {
tok2 = tok2->next()->link();
} else if (Token::Match(tok2, "[{};:] case")) {
const Token *tok3 = Token::findsimplematch(tok2->next(), ": ;", end);
if (!tok3) {
tok2 = tok2->next();
continue;
}
Token::eraseTokens(tok2, tok3->next());
} else if (Token::Match(tok2, "[{};] default : ;")) {
tok2->deleteNext(3);
} else if (tok2->next()->str() == "}") {
--indentlevel2;
if (indentlevel2 <= indentcase)
break;
tok2 = tok2->next();
} else {
tok2 = tok2->next();
}
}
}
break; //stop removing tokens, we arrived to the label.
}
} else if (isgoto && Token::Match(tok, "[{};] do|while|for|BOOST_FOREACH")) {
//it's possible that code inside loop is not dead,
//because of the possible presence of the label pointed by 'goto'
const Token *start = tok->tokAt(2);
if (start && start->str() == "(")
start = start->link()->next();
if (start && start->str() == "{") {
std::string labelpattern = "[{};] " + begin->previous()->str() + " : ;";
bool simplify = true;
for (Token *tok2 = start->next(); tok2 != start->link(); tok2 = tok2->next()) {
if (Token::Match(tok2, labelpattern.c_str())) {
simplify = false;
break;
}
}
//bailout for now
if (!simplify)
break;
}
tok->deleteNext();
} else {
// no need to keep the other strings, remove them.
if (tok->strAt(1) == "while") {
if (tok->str() == "}" && tok->link()->strAt(-1) == "do")
tok->link()->previous()->deleteThis();
}
tok->deleteNext();
}
}
}
//---------------------------------------------------------------------------
void Tokenizer::syntaxError(const Token *tok, const std::string code) const
{
printDebugOutput(0);
throw InternalError(tok, code.empty() ? "syntax error" : "syntax error: " + code, InternalError::SYNTAX);
}
void Tokenizer::unmatchedToken(const Token *tok) const
{
printDebugOutput(0);
throw InternalError(tok,
"Unmatched '" + tok->str() + "'. Configuration: '" + mConfiguration + "'.",
InternalError::SYNTAX);
}
void Tokenizer::syntaxErrorC(const Token *tok, const std::string &what) const
{
printDebugOutput(0);
throw InternalError(tok, "Code '"+what+"' is invalid C code. Use --std or --language to configure the language.", InternalError::SYNTAX);
}
void Tokenizer::unknownMacroError(const Token *tok1) const
{
printDebugOutput(0);
throw InternalError(tok1, "There is an unknown macro here somewhere. Configuration is required. If " + tok1->str() + " is a macro then please configure it.", InternalError::UNKNOWN_MACRO);
}
void Tokenizer::unhandled_macro_class_x_y(const Token *tok) const
{
reportError(tok,
Severity::information,
"class_X_Y",
"The code '" +
tok->str() + " " +
tok->strAt(1) + " " +
tok->strAt(2) + " " +
tok->strAt(3) + "' is not handled. You can use -I or --include to add handling of this code.");
}
void Tokenizer::macroWithSemicolonError(const Token *tok, const std::string &macroName) const
{
reportError(tok,
Severity::information,
"macroWithSemicolon",
"Ensure that '" + macroName + "' is defined either using -I, --include or -D.");
}
void Tokenizer::cppcheckError(const Token *tok) const
{
printDebugOutput(0);
throw InternalError(tok, "Analysis failed. If the code is valid then please report this failure.", InternalError::INTERNAL);
}
/**
* Helper function to check whether number is equal to integer constant X
* or floating point pattern X.0
* @param s the string to check
* @param intConstant the integer constant to check against
* @param floatConstant the string with stringified float constant to check against
* @return true in case s is equal to X or X.0 and false otherwise.
*/
static bool isNumberOneOf(const std::string &s, const MathLib::bigint& intConstant, const char* floatConstant)
{
if (MathLib::isInt(s)) {
if (MathLib::toLongNumber(s) == intConstant)
return true;
} else if (MathLib::isFloat(s)) {
if (MathLib::toString(MathLib::toDoubleNumber(s)) == floatConstant)
return true;
}
return false;
}
// ------------------------------------------------------------------------
// Helper function to check whether number is zero (0 or 0.0 or 0E+0) or not?
// @param s the string to check
// @return true in case s is zero and false otherwise.
// ------------------------------------------------------------------------
bool Tokenizer::isZeroNumber(const std::string &s)
{
return isNumberOneOf(s, 0L, "0.0");
}
// ------------------------------------------------------------------------
// Helper function to check whether number is one (1 or 0.1E+1 or 1E+0) or not?
// @param s the string to check
// @return true in case s is one and false otherwise.
// ------------------------------------------------------------------------
bool Tokenizer::isOneNumber(const std::string &s)
{
if (!MathLib::isPositive(s))
return false;
return isNumberOneOf(s, 1L, "1.0");
}
// ------------------------------------------------------------------------
// Helper function to check whether number is two (2 or 0.2E+1 or 2E+0) or not?
// @param s the string to check
// @return true in case s is two and false otherwise.
// ------------------------------------------------------------------------
bool Tokenizer::isTwoNumber(const std::string &s)
{
if (!MathLib::isPositive(s))
return false;
return isNumberOneOf(s, 2L, "2.0");
}
// ------------------------------------------------------
// Simplify math functions.
// It simplifies the following functions: atol(), fmin(),
// fminl(), fminf(), fmax(), fmaxl(), fmaxf(), pow(),
// powf(), powl(), cbrt(), cbrtl(), cbtrf(), sqrt(),
// sqrtf(), sqrtl(), exp(), expf(), expl(), exp2(),
// exp2f(), exp2l(), log2(), log2f(), log2l(), log1p(),
// log1pf(), log1pl(), log10(), log10l(), log10f(),
// log(), logf(), logl(), logb(), logbf(), logbl(), acosh()
// acoshf(), acoshl(), acos(), acosf(), acosl(), cosh()
// coshf(), coshf(), cos(), cosf(), cosl(), erfc(),
// erfcf(), erfcl(), ilogb(), ilogbf(), ilogbf(), erf(),
// erfl(), erff(), asin(), asinf(), asinf(), asinh(),
// asinhf(), asinhl(), tan(), tanf(), tanl(), tanh(),
// tanhf(), tanhl(), atan(), atanf(), atanl(), atanh(),
// atanhf(), atanhl(), expm1(), expm1l(), expm1f(), sin(),
// sinf(), sinl(), sinh(), sinhf(), sinhl()
// in the tokenlist.
//
// Reference:
// - http://www.cplusplus.com/reference/cmath/
// ------------------------------------------------------
void Tokenizer::simplifyMathFunctions()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->isName() && !tok->varId() && tok->strAt(1) == "(") { // precondition for function
bool simplifcationMade = false;
if (Token::Match(tok, "atol ( %str% )")) { //@todo Add support for atoll()
if (Token::simpleMatch(tok->tokAt(-2), "std ::")) {
tok = tok->tokAt(-2);// set token index two steps back
tok->deleteNext(2); // delete "std ::"
}
const std::string& strNumber = tok->tokAt(2)->strValue(); // get number
const bool isNotAnInteger = (!MathLib::isInt(strNumber));// check: is not an integer
if (strNumber.empty() || isNotAnInteger) {
// Ignore strings which we can't convert
continue;
}
// Convert string into a number and insert into token list
tok->str(MathLib::toString(MathLib::toLongNumber(strNumber)));
// remove ( %num% )
tok->deleteNext(3);
simplifcationMade = true;
} else if (Token::Match(tok, "sqrt|sqrtf|sqrtl|cbrt|cbrtf|cbrtl ( %num% )")) {
// Simplify: sqrt(0) = 0 and cbrt(0) == 0
// sqrt(1) = 1 and cbrt(1) == 1
// get number string
const std::string& parameter(tok->strAt(2));
// is parameter 0 ?
if (isZeroNumber(parameter)) {
tok->deleteNext(3); // delete tokens
tok->str("0"); // insert result into token list
simplifcationMade = true;
} else if (isOneNumber(parameter)) {
tok->deleteNext(3); // delete tokens
tok->str("1"); // insert result into token list
simplifcationMade = true;
}
} else if (Token::Match(tok, "exp|expf|expl|exp2|exp2f|exp2l|cos|cosf|cosl|cosh|coshf|coshl|erfc|erfcf|erfcl ( %num% )")) {
// Simplify: exp[f|l](0) = 1 and exp2[f|l](0) = 1
// cosh[f|l](0) = 1 and cos[f|l](0) = 1
// erfc[f|l](0) = 1
// get number string
const std::string& parameter(tok->strAt(2));
// is parameter 0 ?
if (isZeroNumber(parameter)) {
tok->deleteNext(3); // delete tokens
tok->str("1"); // insert result into token list
simplifcationMade = true;
}
} else if (Token::Match(tok, "log1p|log1pf|log1pl|sin|sinf|sinl|sinh|sinhf|sinhl|erf|erff|erfl|asin|asinf|asinl|asinh|asinhf|asinhl|tan|tanf|tanl|tanh|tanhf|tanhl|atan|atanf|atanl|atanh|atanhf|atanhl|expm1|expm1f|expm1l ( %num% )")) {
// Simplify: log1p[f|l](0) = 0 and sin[f|l](0) = 0
// sinh[f|l](0) = 0 and erf[f|l](0) = 0
// asin[f|l](0) = 0 and sinh[f|l](0) = 0
// tan[f|l](0) = 0 and tanh[f|l](0) = 0
// atan[f|l](0) = 0 and atanh[f|l](0)= 0
// expm1[f|l](0) = 0
// get number string
const std::string& parameter(tok->strAt(2));
// is parameter 0 ?
if (isZeroNumber(parameter)) {
tok->deleteNext(3); // delete tokens
tok->str("0"); // insert result into token list
simplifcationMade = true;
}
} else if (Token::Match(tok, "log2|log2f|log2l|log|logf|logl|log10|log10f|log10l|logb|logbf|logbl|acosh|acoshf|acoshl|acos|acosf|acosl|ilogb|ilogbf|ilogbl ( %num% )")) {
// Simplify: log2[f|l](1) = 0 , log10[f|l](1) = 0
// log[f|l](1) = 0 , logb10[f|l](1) = 0
// acosh[f|l](1) = 0 , acos[f|l](1) = 0
// ilogb[f|l](1) = 0
// get number string
const std::string& parameter(tok->strAt(2));
// is parameter 1 ?
if (isOneNumber(parameter)) {
tok->deleteNext(3); // delete tokens
tok->str("0"); // insert result into token list
simplifcationMade = true;
}
} else if (Token::Match(tok, "fmin|fminl|fminf ( %num% , %num% )")) {
// @todo if one of the parameters is NaN the other is returned
// e.g. printf ("fmin (NaN, -1.0) = %f\n", fmin(NaN,-1.0));
// e.g. printf ("fmin (-1.0, NaN) = %f\n", fmin(-1.0,NaN));
const std::string& strLeftNumber(tok->strAt(2));
const std::string& strRightNumber(tok->strAt(4));
const bool isLessEqual = MathLib::isLessEqual(strLeftNumber, strRightNumber);
// case: left <= right ==> insert left
if (isLessEqual) {
tok->str(strLeftNumber); // insert e.g. -1.0
tok->deleteNext(5); // delete e.g. fmin ( -1.0, 1.0 )
simplifcationMade = true;
} else { // case left > right ==> insert right
tok->str(strRightNumber); // insert e.g. 0.0
tok->deleteNext(5); // delete e.g. fmin ( 1.0, 0.0 )
simplifcationMade = true;
}
} else if (Token::Match(tok, "fmax|fmaxl|fmaxf ( %num% , %num% )")) {
// @todo if one of the parameters is NaN the other is returned
// e.g. printf ("fmax (NaN, -1.0) = %f\n", fmax(NaN,-1.0));
// e.g. printf ("fmax (-1.0, NaN) = %f\n", fmax(-1.0,NaN));
const std::string& strLeftNumber(tok->strAt(2));
const std::string& strRightNumber(tok->strAt(4));
const bool isLessEqual = MathLib::isLessEqual(strLeftNumber, strRightNumber);
// case: left <= right ==> insert right
if (isLessEqual) {
tok->str(strRightNumber);// insert e.g. 1.0
tok->deleteNext(5); // delete e.g. fmax ( -1.0, 1.0 )
simplifcationMade = true;
} else { // case left > right ==> insert left
tok->str(strLeftNumber); // insert e.g. 1.0
tok->deleteNext(5); // delete e.g. fmax ( 1.0, 0.0 )
simplifcationMade = true;
}
} else if (Token::Match(tok, "pow|powf|powl (")) {
if (Token::Match(tok->tokAt(2), "%num% , %num% )")) {
// In case of pow ( 0 , anyNumber > 0): It can be simplified to 0
// In case of pow ( 0 , 0 ): It simplified to 1
// In case of pow ( 1 , anyNumber ): It simplified to 1
const std::string& leftNumber(tok->strAt(2)); // get the left parameter
const std::string& rightNumber(tok->strAt(4)); // get the right parameter
const bool isLeftNumberZero = isZeroNumber(leftNumber);
const bool isLeftNumberOne = isOneNumber(leftNumber);
const bool isRightNumberZero = isZeroNumber(rightNumber);
if (isLeftNumberZero && !isRightNumberZero && MathLib::isPositive(rightNumber)) { // case: 0^(y) = 0 and y > 0
tok->deleteNext(5); // delete tokens
tok->str("0"); // insert simplified result
simplifcationMade = true;
} else if (isLeftNumberZero && isRightNumberZero) { // case: 0^0 = 1
tok->deleteNext(5); // delete tokens
tok->str("1"); // insert simplified result
simplifcationMade = true;
} else if (isLeftNumberOne) { // case 1^(y) = 1
tok->deleteNext(5); // delete tokens
tok->str("1"); // insert simplified result
simplifcationMade = true;
}
}
if (Token::Match(tok->tokAt(2), "%any% , %num% )")) {
// In case of pow( x , 1 ): It can be simplified to x.
const std::string& leftParameter(tok->strAt(2)); // get the left parameter
const std::string& rightNumber(tok->strAt(4)); // get right number
if (isOneNumber(rightNumber)) { // case: x^(1) = x
tok->str(leftParameter); // insert simplified result
tok->deleteNext(5); // delete tokens
simplifcationMade = true;
} else if (isZeroNumber(rightNumber)) { // case: x^(0) = 1
tok->deleteNext(5); // delete tokens
tok->str("1"); // insert simplified result
simplifcationMade = true;
}
}
}
// Jump back to begin of statement if a simplification was performed
if (simplifcationMade) {
while (tok->previous() && tok->str() != ";") {
tok = tok->previous();
}
}
}
}
}
void Tokenizer::simplifyComma()
{
bool inReturn = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
// skip enums
if (Token::Match(tok, "enum class|struct| %name%| :|{")) {
skipEnumBody(&tok);
}
if (!tok)
syntaxError(nullptr); // invalid code like in #4195
if (Token::Match(tok, "(|[") || Token::Match(tok->previous(), "%name%|= {")) {
tok = tok->link();
continue;
}
if (Token::simpleMatch(tok, "= (") && Token::simpleMatch(tok->linkAt(1), ") {")) {
tok = tok->linkAt(1)->linkAt(1);
continue;
}
// Skip unhandled template specifiers..
if (tok->link() && tok->str() == "<")
tok = tok->link();
if (tok->str() == "return" && Token::Match(tok->previous(), "[;{}]"))
inReturn = true;
if (inReturn && Token::Match(tok, "[;{}?:]"))
inReturn = false;
if (!tok->next() || tok->str() != ",")
continue;
// We must not accept just any keyword, e.g. accepting int
// would cause function parameters to corrupt.
if (isCPP() && tok->strAt(1) == "delete") {
// Handle "delete a, delete b;"
tok->str(";");
}
if (isCPP() && Token::Match(tok->tokAt(-2), "delete %name% , %name% ;") &&
tok->next()->varId() != 0) {
// Handle "delete a, b;" - convert to delete a; b;
tok->str(";");
} else if (!inReturn && tok->tokAt(-2)) {
bool replace = false;
for (Token *tok2 = tok->previous(); tok2; tok2 = tok2->previous()) {
if (tok2->str() == "=") {
// Handle "a = 0, b = 0;"
replace = true;
} else if (isCPP() && (Token::Match(tok2, "delete %name%") ||
Token::Match(tok2, "delete [ ] %name%"))) {
// Handle "delete a, a = 0;"
replace = true;
} else if (Token::Match(tok2, "[?:;,{}()]")) {
if (replace && Token::Match(tok2, "[;{}]"))
tok->str(";");
break;
}
}
}
// find token where return ends and also count commas
if (inReturn) {
Token *startFrom = nullptr; // "[;{}]" token before "return"
Token *endAt = nullptr; // first ";" token after "[;{}] return"
// find "; return" pattern before comma
for (Token *tok2 = tok->previous(); tok2; tok2 = tok2->previous()) {
if (tok2->str() == "return") {
startFrom = tok2->previous();
break;
}
}
if (!startFrom)
// to be very sure...
return;
std::size_t commaCounter = 0;
for (Token *tok2 = startFrom->next(); tok2; tok2 = tok2->next()) {
if (tok2->str() == ";") {
endAt = tok2;
break;
} else if (Token::Match(tok2, "(|[") ||
(tok2->str() == "{" && tok2->previous() && tok2->previous()->str() == "=")) {
tok2 = tok2->link();
} else if (tok2->str() == ",") {
++commaCounter;
}
}
if (!endAt)
//probably a syntax error
return;
if (commaCounter) {
// change tokens:
// "; return a ( ) , b ( ) , c ;"
// to
// "; a ( ) ; b ( ) ; return c ;"
// remove "return"
startFrom->deleteNext();
for (Token *tok2 = startFrom->next(); tok2 != endAt; tok2 = tok2->next()) {
if (Token::Match(tok2, "(|[") ||
(tok2->str() == "{" && tok2->previous() && tok2->previous()->str() == "=")) {
tok2 = tok2->link();
} else if (tok2->str() == ",") {
tok2->str(";");
--commaCounter;
if (commaCounter == 0) {
tok2->insertToken("return");
}
}
}
tok = endAt;
}
}
}
}
void Tokenizer::checkConfiguration() const
{
if (!mSettings->checkConfiguration)
return;
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (!Token::Match(tok, "%name% ("))
continue;
if (tok->isControlFlowKeyword())
continue;
for (const Token *tok2 = tok->tokAt(2); tok2 && tok2->str() != ")"; tok2 = tok2->next()) {
if (tok2->str() == ";") {
macroWithSemicolonError(tok, tok->str());
break;
}
if (Token::Match(tok2, "(|{"))
tok2 = tok2->link();
}
}
}
void Tokenizer::validateC() const
{
if (isCPP())
return;
for (const Token *tok = tokens(); tok; tok = tok->next()) {
// That might trigger false positives, but it's much faster to have this truncated pattern
if (Token::Match(tok, "const_cast|dynamic_cast|reinterpret_cast|static_cast <"))
syntaxErrorC(tok, "C++ cast <...");
// Template function..
if (Token::Match(tok, "%name% < %name% > (")) {
const Token *tok2 = tok->tokAt(5);
while (tok2 && !Token::Match(tok2, "[()]"))
tok2 = tok2->next();
if (Token::simpleMatch(tok2, ") {"))
syntaxErrorC(tok, tok->str() + '<' + tok->strAt(2) + ">() {}");
}
if (tok->previous() && !Token::Match(tok->previous(), "[;{}]"))
continue;
if (Token::Match(tok, "using namespace %name% ;"))
syntaxErrorC(tok, "using namespace " + tok->strAt(2));
if (Token::Match(tok, "template < class|typename %name% [,>]"))
syntaxErrorC(tok, "template<...");
if (Token::Match(tok, "%name% :: %name%"))
syntaxErrorC(tok, tok->str() + tok->strAt(1) + tok->strAt(2));
if (Token::Match(tok, "class|namespace %name% [:{]"))
syntaxErrorC(tok, tok->str() + tok->strAt(1) + tok->strAt(2));
}
}
void Tokenizer::validate() const
{
std::stack<const Token *> linkTokens;
const Token *lastTok = nullptr;
for (const Token *tok = tokens(); tok; tok = tok->next()) {
lastTok = tok;
if (Token::Match(tok, "[{([]") || (tok->str() == "<" && tok->link())) {
if (tok->link() == nullptr)
cppcheckError(tok);
linkTokens.push(tok);
}
else if (Token::Match(tok, "[})]]") || (Token::Match(tok, ">|>>") && tok->link())) {
if (tok->link() == nullptr)
cppcheckError(tok);
if (linkTokens.empty() == true)
cppcheckError(tok);
if (tok->link() != linkTokens.top())
cppcheckError(tok);
if (tok != tok->link()->link())
cppcheckError(tok);
linkTokens.pop();
}
else if (tok->link() != nullptr)
cppcheckError(tok);
}
if (!linkTokens.empty())
cppcheckError(linkTokens.top());
// Validate that the Tokenizer::list.back() is updated correctly during simplifications
if (lastTok != list.back())
cppcheckError(lastTok);
}
static const Token *findUnmatchedTernaryOp(const Token * const begin, const Token * const end, unsigned depth = 0)
{
std::stack<const Token *> ternaryOp;
for (const Token *tok = begin; tok != end && tok->str() != ";"; tok = tok->next()) {
if (tok->str() == "?")
ternaryOp.push(tok);
else if (!ternaryOp.empty() && tok->str() == ":")
ternaryOp.pop();
else if (depth < 100 && Token::Match(tok,"(|[")) {
const Token *inner = findUnmatchedTernaryOp(tok->next(), tok->link(), depth+1);
if (inner)
return inner;
tok = tok->link();
}
}
return ternaryOp.empty() ? nullptr : ternaryOp.top();
}
void Tokenizer::findGarbageCode() const
{
// initialization: = {
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (!Token::simpleMatch(tok, "= {"))
continue;
if (Token::simpleMatch(tok->linkAt(1), "} ("))
syntaxError(tok->linkAt(1));
}
// Inside [] there can't be ; or various keywords
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (tok->str() != "[")
continue;
for (const Token *inner = tok->next(); inner != tok->link(); inner = inner->next()) {
if (Token::Match(inner, "(|["))
inner = inner->link();
else if (Token::Match(inner, ";|goto|return|typedef"))
syntaxError(inner);
}
}
// UNKNOWN_MACRO(return)
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (Token::Match(tok, "throw|return )") && Token::Match(tok->linkAt(1)->previous(), "%name% ("))
unknownMacroError(tok->linkAt(1)->previous());
}
// Assign/increment/decrement literal
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (Token::Match(tok, "!!) %num%|%str%|%char% %assign%|++|--"))
syntaxError(tok, tok->next()->str() + " " + tok->strAt(2));
}
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (Token::Match(tok, "if|while|for|switch")) { // if|while|for|switch (EXPR) { ... }
if (tok->previous() && !Token::Match(tok->previous(), "%name%|:|;|{|}|(|)|,"))
syntaxError(tok);
if (Token::Match(tok->previous(), "[(,]"))
continue;
if (!Token::Match(tok->next(), "( !!)"))
syntaxError(tok);
if (tok->str() != "for") {
if (isGarbageExpr(tok->next(), tok->linkAt(1)))
syntaxError(tok);
}
}
}
// keyword keyword
const std::set<std::string> nonConsecutiveKeywords{"break",
"continue",
"for",
"goto",
"if",
"return",
"switch",
"throw",
"typedef",
"while"};
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (!tok->isName() || nonConsecutiveKeywords.count(tok->str()) == 0)
continue;
if (Token::Match(tok, "%name% %name%") && nonConsecutiveKeywords.count(tok->next()->str()) == 1)
syntaxError(tok);
const Token *prev = tok;
while (prev && prev->isName())
prev = prev->previous();
if (Token::Match(prev, "%op%|%num%|%str%|%char%"))
syntaxError(tok, prev == tok->previous() ? (prev->str() + " " + tok->str()) : (prev->str() + " .. " + tok->str()));
}
// case keyword must be inside switch
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "switch (")) {
if (Token::simpleMatch(tok->linkAt(1), ") {")) {
tok = tok->linkAt(1)->linkAt(1);
continue;
}
const Token *switchToken = tok;
tok = tok->linkAt(1);
if (!tok)
syntaxError(switchToken);
// Look for the end of the switch statement, i.e. the first semi-colon or '}'
for (; tok ; tok = tok->next()) {
if (tok->str() == "{") {
tok = tok->link();
}
if (Token::Match(tok, ";|}")) {
// We're at the end of the switch block
if (tok->str() == "}" && tok->strAt(-1) == ":") // Invalid case
syntaxError(switchToken);
break;
}
}
if (!tok)
break;
} else if (tok->str() == "(") {
tok = tok->link();
} else if (tok->str() == "case") {
syntaxError(tok);
}
}
for (const Token *tok = tokens(); tok ; tok = tok->next()) {
if (!Token::simpleMatch(tok, "for (")) // find for loops
continue;
// count number of semicolons
unsigned int semicolons = 0;
const Token* const startTok = tok;
tok = tok->next()->link()->previous(); // find ")" of the for-loop
// walk backwards until we find the beginning (startTok) of the for() again
for (; tok != startTok; tok = tok->previous()) {
if (tok->str() == ";") { // do the counting
semicolons++;
} else if (tok->str() == ")") { // skip pairs of ( )
tok = tok->link();
}
}
// if we have an invalid number of semicolons inside for( ), assume syntax error
if ((semicolons == 1) || (semicolons > 2)) {
syntaxError(tok);
}
}
// Operators without operands..
const Token *templateEndToken = nullptr;
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (!templateEndToken) {
if (tok->str() == "<" && isCPP())
templateEndToken = tok->findClosingBracket();
} else {
if (templateEndToken == tok)
templateEndToken = nullptr;
if (Token::Match(tok, "> %cop%"))
continue;
}
if (Token::Match(tok, "%or%|%oror%|==|!=|+|-|/|!|>=|<=|~|++|--|::|sizeof|throw|decltype|typeof {|if|else|try|catch|while|do|for|return|switch|break|namespace"))
syntaxError(tok);
if (Token::Match(tok, "( %any% )") && tok->next()->isKeyword() && !Token::simpleMatch(tok->next(), "void"))
syntaxError(tok);
if (Token::Match(tok, "%num%|%bool%|%char%|%str% %num%|%bool%|%char%|%str%") && !Token::Match(tok, "%str% %str%"))
syntaxError(tok);
if (Token::Match(tok, "%assign% typename|class %assign%"))
syntaxError(tok);
if (Token::Match(tok, "%cop%|=|,|[ %or%|%oror%|/|%"))
syntaxError(tok);
if (Token::Match(tok, ";|(|[ %comp%"))
syntaxError(tok);
if (Token::Match(tok, "%cop%|= ]") && !(isCPP() && Token::Match(tok->previous(), "[|, &|= ]")))
syntaxError(tok);
if (Token::Match(tok, "[+-] [;,)]}]"))
syntaxError(tok);
}
// ternary operator without :
if (const Token *ternaryOp = findUnmatchedTernaryOp(tokens(), nullptr))
syntaxError(ternaryOp);
// Code must not start with an arithmetical operand
if (Token::Match(list.front(), "%cop%"))
syntaxError(list.front());
// Code must end with } ; ) NAME
if (!Token::Match(list.back(), "%name%|;|}|)"))
syntaxError(list.back());
if (list.back()->str() == ")" && !Token::Match(list.back()->link()->previous(), "%name% ("))
syntaxError(list.back());
for (const Token *end = list.back(); end && end->isName(); end = end->previous()) {
if (Token::Match(end, "void|char|short|int|long|float|double|const|volatile|static|inline|struct|class|enum|union|template|sizeof|case|break|continue|typedef"))
syntaxError(list.back());
}
if ((list.back()->str()==")" || list.back()->str()=="}") && list.back()->previous() && list.back()->previous()->isControlFlowKeyword())
syntaxError(list.back()->previous());
// Garbage templates..
if (isCPP()) {
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (!Token::simpleMatch(tok, "template <"))
continue;
if (tok->previous() && !Token::Match(tok->previous(), "[:;{})>]"))
syntaxError(tok);
const Token * const tok1 = tok;
tok = tok->next()->findClosingBracket();
if (!tok)
syntaxError(tok1);
if (!Token::Match(tok, ">|>> ::| %name%") &&
!Token::Match(tok, ">|>> [ [ %name%"))
syntaxError(tok->next() ? tok->next() : tok1);
}
}
// Objective C/C++
for (const Token *tok = tokens(); tok; tok = tok->next()) {
if (Token::Match(tok, "[;{}] [ %name% %name% ] ;"))
syntaxError(tok->next());
}
}
bool Tokenizer::isGarbageExpr(const Token *start, const Token *end)
{
for (const Token *tok = start; tok != end; tok = tok->next()) {
if (tok->isControlFlowKeyword())
return true;
if (tok->str() == ";")
return true;
if (tok->str() == "{")
tok = tok->link();
}
return false;
}
std::string Tokenizer::simplifyString(const std::string &source)
{
std::string str = source;
for (std::string::size_type i = 0; i + 1U < str.size(); ++i) {
if (str[i] != '\\')
continue;
int c = 'a'; // char
unsigned int sz = 0; // size of stringdata
if (str[i+1] == 'x') {
sz = 2;
while (sz < 4 && std::isxdigit((unsigned char)str[i+sz]))
sz++;
if (sz > 2) {
std::istringstream istr(str.substr(i+2, sz-2));
istr >> std::hex >> c;
}
} else if (MathLib::isOctalDigit(str[i+1])) {
sz = 2;
while (sz < 4 && MathLib::isOctalDigit(str[i+sz]))
sz++;
std::istringstream istr(str.substr(i+1, sz-1));
istr >> std::oct >> c;
str = str.substr(0,i) + (char)c + str.substr(i+sz);
continue;
}
if (sz <= 2)
i++;
else if (i+sz < str.size())
str.replace(i, sz, std::string(1U, (char)c));
else
str.replace(i, str.size() - i - 1U, "a");
}
return str;
}
void Tokenizer::simplifyWhile0()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
// while (0)
const bool while0(Token::Match(tok->previous(), "[{};] while ( 0|false )"));
// for (0) - not banal, ticket #3140
const bool for0((Token::Match(tok->previous(), "[{};] for ( %name% = %num% ; %name% < %num% ;") &&
tok->strAt(2) == tok->strAt(6) && tok->strAt(4) == tok->strAt(8)) ||
(Token::Match(tok->previous(), "[{};] for ( %type% %name% = %num% ; %name% < %num% ;") &&
tok->strAt(3) == tok->strAt(7) && tok->strAt(5) == tok->strAt(9)));
if (!while0 && !for0)
continue;
if (while0 && tok->previous()->str() == "}") {
// find "do"
Token *tok2 = tok->previous()->link();
tok2 = tok2->previous();
if (tok2 && tok2->str() == "do") {
const bool flowmatch = Token::findmatch(tok2, "continue|break", tok) != nullptr;
// delete "do ({)"
tok2->deleteThis();
if (!flowmatch)
tok2->deleteThis();
// delete "(}) while ( 0 ) (;)"
tok = tok->previous();
tok->deleteNext(4); // while ( 0 )
if (tok->next() && tok->next()->str() == ";")
tok->deleteNext(); // ;
if (!flowmatch)
tok->deleteThis(); // }
continue;
}
}
// remove "while (0) { .. }"
if (Token::simpleMatch(tok->next()->link(), ") {")) {
Token *end = tok->next()->link(), *old_prev = tok->previous();
end = end->next()->link();
if (Token::Match(tok, "for ( %name% ="))
old_prev = end->link();
eraseDeadCode(old_prev, end->next());
if (old_prev && old_prev->next())
tok = old_prev->next();
else
break;
}
}
}
void Tokenizer::simplifyErrNoInWhile()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() != "errno")
continue;
Token *endpar = nullptr;
if (Token::Match(tok->previous(), "&& errno == EINTR ) { ;| }"))
endpar = tok->tokAt(3);
else if (Token::Match(tok->tokAt(-2), "&& ( errno == EINTR ) ) { ;| }"))
endpar = tok->tokAt(4);
else
continue;
if (Token::simpleMatch(endpar->link()->previous(), "while (")) {
Token *tok1 = tok->previous();
if (tok1->str() == "(")
tok1 = tok1->previous();
// erase "&& errno == EINTR"
tok1 = tok1->previous();
Token::eraseTokens(tok1, endpar);
// tok is invalid.. move to endpar
tok = endpar;
}
}
}
void Tokenizer::simplifyFuncInWhile()
{
unsigned int count = 0;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!Token::Match(tok, "while ( %name% ( %name% ) ) {"))
continue;
Token *func = tok->tokAt(2);
const Token * const var = tok->tokAt(4);
Token * const end = tok->next()->link()->next()->link();
const unsigned int varid = ++mVarId; // Create new variable
const std::string varname("cppcheck:r" + MathLib::toString(++count));
tok->str("int");
tok->next()->insertToken(varname);
tok->tokAt(2)->varId(varid);
tok->insertToken("while");
tok->insertToken(";");
tok->insertToken(")");
tok->insertToken(var->str());
tok->next()->varId(var->varId());
tok->insertToken("(");
tok->insertToken(func->str());
tok->insertToken("=");
tok->insertToken(varname);
tok->next()->varId(varid);
Token::createMutualLinks(tok->tokAt(4), tok->tokAt(6));
end->previous()->insertToken(varname);
end->previous()->varId(varid);
end->previous()->insertToken("=");
Token::move(func, func->tokAt(3), end->previous());
end->previous()->insertToken(";");
tok = end;
}
}
void Tokenizer::simplifyStructDecl()
{
const bool cpp = isCPP();
// A counter that is used when giving unique names for anonymous structs.
unsigned int count = 0;
// Skip simplification of unions in class definition
std::stack<bool> skip; // true = in function, false = not in function
skip.push(false);
// Add names for anonymous structs
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName())
continue;
// check for anonymous struct/union
if (Token::Match(tok, "struct|union {")) {
if (Token::Match(tok->next()->link(), "} const| *|&| const| %type% ,|;|[|(|{|=")) {
tok->insertToken("Anonymous" + MathLib::toString(count++));
}
}
// check for derived anonymous class/struct
else if (cpp && Token::Match(tok, "class|struct :")) {
const Token *tok1 = Token::findsimplematch(tok, "{");
if (tok1 && Token::Match(tok1->link(), "} const| *|&| const| %type% ,|;|[|(|{")) {
tok->insertToken("Anonymous" + MathLib::toString(count++));
}
}
// check for anonymous enum
else if ((Token::simpleMatch(tok, "enum {") &&
!Token::Match(tok->tokAt(-3), "using %name% =") &&
Token::Match(tok->next()->link(), "} %type%| ,|;|[|(|{")) ||
(Token::Match(tok, "enum : %type% {") && Token::Match(tok->linkAt(3), "} %type%| ,|;|[|(|{"))) {
tok->insertToken("Anonymous" + MathLib::toString(count++));
}
}
for (Token *tok = list.front(); tok; tok = tok->next()) {
// check for start of scope and determine if it is in a function
if (tok->str() == "{")
skip.push(Token::Match(tok->previous(), "const|)"));
// end of scope
else if (tok->str() == "}" && !skip.empty())
skip.pop();
// check for named struct/union
else if (Token::Match(tok, "class|struct|union|enum %type% :|{")) {
Token *start = tok;
while (Token::Match(start->previous(), "%type%"))
start = start->previous();
const Token * const type = tok->next();
Token *next = tok->tokAt(2);
while (next && next->str() != "{")
next = next->next();
if (!next)
continue;
skip.push(false);
tok = next->link();
if (!tok)
break; // see #4869 segmentation fault in Tokenizer::simplifyStructDecl (invalid code)
Token *restart = next;
// check for named type
if (Token::Match(tok->next(), "const| *|&| const| %type% ,|;|[|=|(|{")) {
tok->insertToken(";");
tok = tok->next();
while (!Token::Match(start, "struct|class|union|enum")) {
tok->insertToken(start->str());
tok = tok->next();
start->deleteThis();
}
if (!tok)
break; // see #4869 segmentation fault in Tokenizer::simplifyStructDecl (invalid code)
tok->insertToken(type->str());
if (start->str() != "class") {
tok->insertToken(start->str());
tok = tok->next();
}
tok = tok->tokAt(2);
// check for initialization
if (tok && (tok->next()->str() == "(" || tok->next()->str() == "{")) {
tok->insertToken("=");
tok = tok->next();
if (start->str() == "enum") {
if (tok->next()->str() == "{") {
tok->next()->str("(");
tok->linkAt(1)->str(")");
}
}
}
}
tok = restart;
}
// check for anonymous struct/union
else if (Token::Match(tok, "struct|union {")) {
const bool inFunction = skip.top();
skip.push(false);
Token *tok1 = tok;
Token *restart = tok->next();
tok = tok->next()->link();
// unnamed anonymous struct/union so possibly remove it
if (tok && tok->next() && tok->next()->str() == ";") {
if (inFunction && tok1->str() == "union") {
// Try to create references in the union..
Token *tok2 = tok1->tokAt(2);
while (tok2) {
if (Token::Match(tok2, "%type% %name% ;"))
tok2 = tok2->tokAt(3);
else
break;
}
if (!Token::simpleMatch(tok2, "} ;"))
continue;
Token *vartok = nullptr;
tok2 = tok1->tokAt(2);
while (Token::Match(tok2, "%type% %name% ;")) {
if (!vartok) {
vartok = tok2->next();
tok2 = tok2->tokAt(3);
} else {
tok2->insertToken("&");
tok2 = tok2->tokAt(2);
tok2->insertToken(vartok->str());
tok2->next()->varId(vartok->varId());
tok2->insertToken("=");
tok2 = tok2->tokAt(4);
}
}
}
// don't remove unnamed anonymous unions from a class, struct or union
if (!(!inFunction && tok1->str() == "union") && !Token::Match(tok1->tokAt(-3), "using %name% =")) {
skip.pop();
tok1->deleteThis();
if (tok1->next() == tok) {
tok1->deleteThis();
tok = tok1;
} else
tok1->deleteThis();
restart = tok1->previous();
tok->deleteThis();
if (tok->next())
tok->deleteThis();
}
}
if (!restart) {
simplifyStructDecl();
return;
} else if (!restart->next())
return;
tok = restart;
}
}
}
void Tokenizer::simplifyCallingConvention()
{
const bool windows = mSettings->isWindowsPlatform();
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (Token::Match(tok, "__cdecl|__stdcall|__fastcall|__thiscall|__clrcall|__syscall|__pascal|__fortran|__far|__near") || (windows && Token::Match(tok, "WINAPI|APIENTRY|CALLBACK"))) {
tok->deleteThis();
}
}
}
void Tokenizer::simplifyDeclspec()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
while (Token::Match(tok, "__declspec|_declspec (") && tok->next()->link() && tok->next()->link()->next()) {
if (Token::Match(tok->tokAt(2), "noreturn|nothrow")) {
Token *tok1 = tok->next()->link()->next();
while (tok1 && !Token::Match(tok1, "%name%")) {
tok1 = tok1->next();
}
if (tok1) {
if (tok->strAt(2) == "noreturn")
tok1->isAttributeNoreturn(true);
else
tok1->isAttributeNothrow(true);
}
} else if (tok->strAt(2) == "property")
tok->next()->link()->insertToken("__property");
Token::eraseTokens(tok, tok->next()->link()->next());
tok->deleteThis();
}
}
}
void Tokenizer::simplifyAttribute()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%type% (") && !mSettings->library.isNotLibraryFunction(tok)) {
if (mSettings->library.isFunctionConst(tok->str(), true))
tok->isAttributePure(true);
if (mSettings->library.isFunctionConst(tok->str(), false))
tok->isAttributeConst(true);
}
while (Token::Match(tok, "__attribute__|__attribute (") && tok->next()->link() && tok->next()->link()->next()) {
if (Token::Match(tok->tokAt(2), "( constructor|__constructor__")) {
// prototype for constructor is: void func(void);
if (!tok->next()->link()->next())
syntaxError(tok);
if (tok->next()->link()->next()->str() == "void") { // __attribute__((constructor)) void func() {}
if (!tok->next()->link()->next()->next())
syntaxError(tok);
tok->next()->link()->next()->next()->isAttributeConstructor(true);
} else if (tok->next()->link()->next()->str() == ";" && tok->linkAt(-1) && tok->previous()->link()->previous()) // void func() __attribute__((constructor));
tok->previous()->link()->previous()->isAttributeConstructor(true);
else // void __attribute__((constructor)) func() {}
tok->next()->link()->next()->isAttributeConstructor(true);
}
else if (Token::Match(tok->tokAt(2), "( destructor|__destructor__")) {
// prototype for destructor is: void func(void);
if (!tok->next()->link()->next())
syntaxError(tok);
if (tok->next()->link()->next()->str() == "void") // __attribute__((destructor)) void func() {}
tok->next()->link()->next()->next()->isAttributeDestructor(true);
else if (tok->next()->link()->next()->str() == ";" && tok->linkAt(-1) && tok->previous()->link()->previous()) // void func() __attribute__((destructor));
tok->previous()->link()->previous()->isAttributeDestructor(true);
else // void __attribute__((destructor)) func() {}
tok->next()->link()->next()->isAttributeDestructor(true);
}
else if (Token::Match(tok->tokAt(2), "( unused|__unused__|used|__used__ )")) {
Token *vartok = nullptr;
// check if after variable name
if (Token::Match(tok->next()->link()->next(), ";|=")) {
if (Token::Match(tok->previous(), "%type%"))
vartok = tok->previous();
}
// check if before variable name
else if (Token::Match(tok->next()->link()->next(), "%type%"))
vartok = tok->next()->link()->next();
if (vartok) {
const std::string &attribute(tok->strAt(3));
if (attribute.find("unused") != std::string::npos)
vartok->isAttributeUnused(true);
else
vartok->isAttributeUsed(true);
}
}
else if (Token::Match(tok->tokAt(2), "( pure|__pure__|const|__const__|noreturn|__noreturn__|nothrow|__nothrow__|warn_unused_result )")) {
Token *functok = nullptr;
// type func(...) __attribute__((attribute));
if (tok->previous() && tok->previous()->link() && Token::Match(tok->previous()->link()->previous(), "%name% ("))
functok = tok->previous()->link()->previous();
// type __attribute__((attribute)) func() { }
else {
Token *tok2 = tok->next()->link();
while (Token::Match(tok2, ") __attribute__|__attribute ("))
tok2 = tok2->linkAt(2);
if (Token::Match(tok2, ") %name% ("))
functok = tok2->next();
}
if (functok) {
const std::string &attribute(tok->strAt(3));
if (attribute.find("pure") != std::string::npos)
functok->isAttributePure(true);
else if (attribute.find("const") != std::string::npos)
functok->isAttributeConst(true);
else if (attribute.find("noreturn") != std::string::npos)
functok->isAttributeNoreturn(true);
else if (attribute.find("nothrow") != std::string::npos)
functok->isAttributeNothrow(true);
else if (attribute.find("warn_unused_result") != std::string::npos)
functok->isAttributeNodiscard(true);
}
}
else if (Token::simpleMatch(tok->previous(), "} __attribute__ ( ( packed )")) {
tok->previous()->isAttributePacked(true);
}
Token::eraseTokens(tok, tok->next()->link()->next());
tok->deleteThis();
}
}
}
void Tokenizer::simplifyCPPAttribute()
{
if (mSettings->standards.cpp < Standards::CPP11 || isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->link() || !Token::Match(tok, "[ [ %name%"))
continue;
if (Token::Match(tok->tokAt(2), "noreturn|nodiscard")) {
const Token * head = tok->link()->next();
while (Token::Match(head, "%name%|::|*|&"))
head = head->next();
if (head && head->str() == "(" && isFunctionHead(head, "{|;")) {
if (tok->strAt(2) == "noreturn")
head->previous()->isAttributeNoreturn(true);
else
head->previous()->isAttributeNodiscard(true);
}
}
Token::eraseTokens(tok, tok->link()->next());
tok->deleteThis();
}
}
static const std::set<std::string> keywords = {
"inline"
, "_inline"
, "__inline"
, "__forceinline"
, "register"
, "__restrict"
, "__restrict__"
, "__thread"
};
// Remove "inline", "register", "restrict", "override", "static" and "constexpr"
// "restrict" keyword
// - New to 1999 ANSI/ISO C standard
// - Not in C++ standard yet
void Tokenizer::simplifyKeyword()
{
// FIXME: There is a risk that "keywords" are removed by mistake. This
// code should be fixed so it doesn't remove variables etc. Nonstandard
// keywords should be defined with a library instead. For instance the
// linux kernel code at least uses "_inline" as struct member name at some
// places.
const bool c99 = isC() && mSettings->standards.c >= Standards::C99;
const bool cpp11 = isCPP() && mSettings->standards.cpp >= Standards::CPP11;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (keywords.find(tok->str()) != keywords.end()) {
// Don't remove struct members
if (!Token::simpleMatch(tok->previous(), "."))
tok->deleteThis(); // Simplify..
}
if (isC() || mSettings->standards.cpp == Standards::CPP03) {
if (tok->str() == "auto")
tok->deleteThis();
}
// simplify static keyword:
// void foo( int [ static 5 ] ); ==> void foo( int [ 5 ] );
if (Token::Match(tok, "[ static %num%"))
tok->deleteNext();
if (c99) {
while (tok->str() == "restrict")
tok->deleteThis();
if (mSettings->standards.c >= Standards::C11) {
while (tok->str() == "_Atomic")
tok->deleteThis();
}
}
else if (cpp11) {
if (tok->str() == "constexpr") {
tok->originalName(tok->str());
tok->str("const");
}
// final:
// 1) struct name final { }; <- struct is final
if (Token::Match(tok->previous(), "struct|class|union %type% final [:{]")) {
tok->deleteNext();
}
// noexcept -> noexcept(true)
// 2) void f() noexcept; -> void f() noexcept(true);
else if (Token::Match(tok, ") noexcept :|{|;|const|override|final")) {
// Insertion is done in inverse order
// The brackets are linked together accordingly afterwards
Token * tokNoExcept = tok->next();
tokNoExcept->insertToken(")");
Token * braceEnd = tokNoExcept->next();
tokNoExcept->insertToken("true");
tokNoExcept->insertToken("(");
Token * braceStart = tokNoExcept->next();
tok = tok->tokAt(3);
Token::createMutualLinks(braceStart, braceEnd);
}
// 3) thread_local -> static
// on single thread thread_local has the effect of static
else if (tok->str() == "thread_local") {
tok->originalName(tok->str());
tok->str("static");
}
}
}
}
void Tokenizer::simplifyAssignmentInFunctionCall()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "(")
tok = tok->link();
// Find 'foo(var='. Exclude 'assert(var=' to allow tests to check that assert(...) does not contain side-effects
else if (Token::Match(tok, "[;{}] %name% ( %name% =") &&
Token::simpleMatch(tok->linkAt(2), ") ;") &&
!Token::Match(tok->next(), "assert|while")) {
const std::string& funcname(tok->next()->str());
Token* const vartok = tok->tokAt(3);
// Goto ',' or ')'..
for (Token *tok2 = vartok->tokAt(2); tok2; tok2 = tok2->next()) {
if (tok2->link() && Token::Match(tok2, "(|[|{"))
tok2 = tok2->link();
else if (tok2->str() == ";")
break;
else if (Token::Match(tok2, ")|,")) {
tok2 = tok2->previous();
tok2->insertToken(vartok->str());
tok2->next()->varId(vartok->varId());
tok2->insertToken("(");
Token::createMutualLinks(tok2->next(), tok->linkAt(2));
tok2->insertToken(funcname);
tok2->insertToken(";");
Token::eraseTokens(tok, vartok);
break;
}
}
}
}
}
void Tokenizer::simplifyAssignmentBlock()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "[;{}] %name% = ( {")) {
const std::string &varname = tok->next()->str();
// goto the "} )"
unsigned int indentlevel = 0;
Token *tok2 = tok;
while (nullptr != (tok2 = tok2->next())) {
if (Token::Match(tok2, "(|{"))
++indentlevel;
else if (Token::Match(tok2, ")|}")) {
if (indentlevel <= 2)
break;
--indentlevel;
} else if (indentlevel == 2 && tok2->str() == varname && Token::Match(tok2->previous(), "%type%|*"))
// declaring variable in inner scope with same name as lhs variable
break;
}
if (indentlevel == 2 && Token::simpleMatch(tok2, "} )")) {
tok2 = tok2->tokAt(-3);
if (Token::Match(tok2, "[;{}] %num%|%name% ;")) {
tok2->insertToken("=");
tok2->insertToken(tok->next()->str());
tok2->next()->varId(tok->next()->varId());
tok->deleteNext(3);
tok2->tokAt(5)->deleteNext();
}
}
}
}
}
// Remove __asm..
void Tokenizer::simplifyAsm()
{
std::string instruction;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "__asm|_asm|asm {") &&
tok->next()->link()->next()) {
instruction = tok->tokAt(2)->stringifyList(tok->next()->link());
Token::eraseTokens(tok, tok->next()->link()->next());
}
else if (Token::Match(tok, "asm|__asm|__asm__ volatile|__volatile|__volatile__| (")) {
// Goto "("
Token *partok = tok->next();
if (partok->str() != "(")
partok = partok->next();
instruction = partok->next()->stringifyList(partok->link());
Token::eraseTokens(tok, partok->link()->next());
}
else if (Token::Match(tok, "_asm|__asm")) {
Token *endasm = tok->next();
const Token *firstSemiColon = nullptr;
unsigned int comment = 0;
while (Token::Match(endasm, "%num%|%name%|,|:|;") || (endasm && endasm->linenr() == comment)) {
if (Token::Match(endasm, "_asm|__asm|__endasm"))
break;
if (endasm->str() == ";") {
comment = endasm->linenr();
if (!firstSemiColon)
firstSemiColon = endasm;
}
endasm = endasm->next();
}
if (Token::simpleMatch(endasm, "__endasm")) {
instruction = tok->next()->stringifyList(endasm);
Token::eraseTokens(tok, endasm->next());
if (!Token::simpleMatch(tok->next(), ";"))
tok->insertToken(";");
} else if (firstSemiColon) {
instruction = tok->next()->stringifyList(firstSemiColon);
Token::eraseTokens(tok, firstSemiColon);
} else if (!endasm) {
instruction = tok->next()->stringifyList(endasm);
Token::eraseTokens(tok, endasm);
tok->insertToken(";");
} else
continue;
}
else
continue;
// insert "asm ( "instruction" )"
tok->str("asm");
if (tok->strAt(1) != ";" && tok->strAt(1) != "{")
tok->insertToken(";");
tok->insertToken(")");
tok->insertToken("\"" + instruction + "\"");
tok->insertToken("(");
tok = tok->next();
Token::createMutualLinks(tok, tok->tokAt(2));
//move the new tokens in the same line as ";" if available
tok = tok->tokAt(2);
if (tok->next() && tok->next()->str() == ";" &&
tok->next()->linenr() != tok->linenr()) {
const unsigned int endposition = tok->next()->linenr();
tok = tok->tokAt(-3);
for (int i = 0; i < 4; ++i) {
tok = tok->next();
tok->linenr(endposition);
}
}
}
}
void Tokenizer::simplifyAsm2()
{
// Block declarations: ^{}
// A C extension used to create lambda like closures.
// Put ^{} statements in asm()
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() != "^")
continue;
if (Token::simpleMatch(tok, "^ {") || (Token::simpleMatch(tok->linkAt(1), ") {") && tok->strAt(-1) != "operator")) {
Token * start = tok;
while (start && !Token::Match(start, "[,(;{}=]")) {
if (start->link() && Token::Match(start, ")|]|>"))
start = start->link();
start = start->previous();
}
const Token *last = tok->next()->link();
if (Token::simpleMatch(last, ") {"))
last = last->linkAt(1);
last = last->next();
while (last && !Token::Match(last, "%cop%|,|;|{|}|)")) {
if (Token::Match(last, "(|["))
last = last->link();
last = last->next();
}
if (start && last) {
std::string asmcode;
while (start->next() != last) {
asmcode += start->next()->str();
start->deleteNext();
}
if (last->str() == "}")
start->insertToken(";");
start->insertToken(")");
start->insertToken("\"" + asmcode + "\"");
start->insertToken("(");
start->insertToken("asm");
start->tokAt(2)->link(start->tokAt(4));
start->tokAt(4)->link(start->tokAt(2));
tok = start->tokAt(4);
}
}
}
}
void Tokenizer::simplifyAt()
{
std::set<std::string> var;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "%name% @ %num% ;")) {
var.insert(tok->str());
tok->isAtAddress(true);
Token::eraseTokens(tok,tok->tokAt(3));
}
if (Token::Match(tok, "%name% @ %num% : %num% ;")) {
var.insert(tok->str());
tok->isAtAddress(true);
Token::eraseTokens(tok,tok->tokAt(5));
}
if (Token::Match(tok, "%name% @ %name% : %num% ;") && var.find(tok->strAt(2)) != var.end()) {
var.insert(tok->str());
tok->isAtAddress(true);
Token::eraseTokens(tok,tok->tokAt(5));
}
// array declaration
if (Token::Match(tok, "] @ %num% ;")) {
tok->isAtAddress(true);
Token::eraseTokens(tok,tok->tokAt(3));
}
// keywords in compiler from cosmic software for STM8
// TODO: Should use platform configuration.
if (Token::Match(tok, "@ builtin|eeprom|far|inline|interrupt|near|noprd|nostack|nosvf|packed|stack|svlreg|tiny|vector")) {
tok->str(tok->next()->str() + "@");
tok->deleteNext();
}
}
}
// Simplify bitfields
void Tokenizer::simplifyBitfields()
{
bool goback = false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (goback) {
goback = false;
tok = tok->previous();
}
Token *last = nullptr;
if (!Token::Match(tok, ";|{|}|public:|protected:|private:"))
continue;
bool isEnum = false;
if (tok->str() == "}") {
const Token *type = tok->link()->previous();
while (type && type->isName()) {
if (type->str() == "enum") {
isEnum = true;
break;
}
type = type->previous();
}
}
if (Token::Match(tok->next(), "const| %type% %name% :") &&
!Token::Match(tok->next(), "case|public|protected|private|class|struct") &&
!Token::simpleMatch(tok->tokAt(2), "default :")) {
Token *tok1 = (tok->next()->str() == "const") ? tok->tokAt(3) : tok->tokAt(2);
if (Token::Match(tok1, "%name% : %num% ;"))
tok1->setBits(MathLib::toLongNumber(tok1->strAt(2)));
if (tok1 && tok1->tokAt(2) &&
(Token::Match(tok1->tokAt(2), "%bool%|%num%") ||
!Token::Match(tok1->tokAt(2), "public|protected|private| %type% ::|<|,|{|;"))) {
while (tok1->next() && !Token::Match(tok1->next(), "[;,)]{}]")) {
if (Token::Match(tok1->next(), "[([]"))
Token::eraseTokens(tok1, tok1->next()->link());
tok1->deleteNext();
}
last = tok1->next();
}
} else if (isEnum && Token::Match(tok, "} %name%| : %num% ;")) {
if (tok->next()->str() == ":") {
tok->deleteNext(2);
tok->insertToken("Anonymous");
} else {
tok->next()->deleteNext(2);
}
} else if (Token::Match(tok->next(), "const| %type% : %num%|%bool% ;") &&
tok->next()->str() != "default") {
const int offset = (tok->next()->str() == "const") ? 1 : 0;
if (!Token::Match(tok->tokAt(3 + offset), "[{};()]")) {
tok->deleteNext(4 + offset);
goback = true;
}
}
if (last && last->str() == ",") {
Token * tok1 = last;
tok1->str(";");
const Token *const tok2 = tok->next();
tok1->insertToken(tok2->str());
tok1 = tok1->next();
tok1->isSigned(tok2->isSigned());
tok1->isUnsigned(tok2->isUnsigned());
tok1->isLong(tok2->isLong());
}
}
}
// Types and objects in std namespace that are neither functions nor templates
static const std::set<std::string> stdTypes = {
"string", "wstring", "u16string", "u32string",
"iostream", "ostream", "ofstream", "ostringstream",
"istream", "ifstream", "istringstream", "fstream", "stringstream",
"wstringstream", "wistringstream", "wostringstream", "wstringbuf",
"stringbuf", "streambuf", "ios", "filebuf", "ios_base",
"exception", "bad_exception", "bad_alloc",
"logic_error", "domain_error", "invalid_argument_", "length_error",
"out_of_range", "runtime_error", "range_error", "overflow_error", "underflow_error",
"locale",
"cout", "cerr", "clog", "cin",
"wcerr", "wcin", "wclog", "wcout",
"endl", "ends", "flush",
"boolalpha", "noboolalpha", "showbase", "noshowbase",
"showpoint", "noshowpoint", "showpos", "noshowpos",
"skipws", "noskipws", "unitbuf", "nounitbuf", "uppercase", "nouppercase",
"dec", "hex", "oct",
"fixed", "scientific",
"internal", "left", "right",
"fpos", "streamoff", "streampos", "streamsize"
};
static const std::set<std::string> stdTemplates = {
"array", "basic_string", "bitset", "deque", "list", "map", "multimap",
"priority_queue", "queue", "set", "multiset", "stack", "vector", "pair",
"iterator", "iterator_traits",
"unordered_map", "unordered_multimap", "unordered_set", "unordered_multiset",
"tuple", "function"
};
static const std::set<std::string> stdFunctions = {
"getline",
"for_each", "find", "find_if", "find_end", "find_first_of",
"adjacent_find", "count", "count_if", "mismatch", "equal", "search", "search_n",
"copy", "copy_backward", "swap", "swap_ranges", "iter_swap", "transform", "replace",
"replace_if", "replace_copy", "replace_copy_if", "fill", "fill_n", "generate", "generate_n", "remove",
"remove_if", "remove_copy", "remove_copy_if",
"unique", "unique_copy", "reverse", "reverse_copy",
"rotate", "rotate_copy", "random_shuffle", "partition", "stable_partition",
"sort", "stable_sort", "partial_sort", "partial_sort_copy", "nth_element",
"lower_bound", "upper_bound", "equal_range", "binary_search", "merge", "inplace_merge", "includes",
"set_union", "set_intersection", "set_difference",
"set_symmetric_difference", "push_heap", "pop_heap", "make_heap", "sort_heap",
"min", "max", "min_element", "max_element", "lexicographical_compare", "next_permutation", "prev_permutation",
"advance", "back_inserter", "distance", "front_inserter", "inserter",
"make_pair", "make_shared", "make_tuple"
};
// Add std:: in front of std classes, when using namespace std; was given
void Tokenizer::simplifyNamespaceStd()
{
if (!isCPP())
return;
const bool isCPP11 = mSettings->standards.cpp == Standards::CPP11;
std::set<std::string> userFunctions;
for (const Token* tok = Token::findsimplematch(list.front(), "using namespace std ;"); tok; tok = tok->next()) {
bool insert = false;
if (Token::Match(tok, "enum class|struct| %name%| :|{")) { // Don't replace within enum definitions
skipEnumBody(&tok);
}
if (!Token::Match(tok->previous(), ".|::")) {
if (Token::Match(tok, "%name% (")) {
if (isFunctionHead(tok->next(), "{"))
userFunctions.insert(tok->str());
else if (isFunctionHead(tok->next(), ";")) {
const Token *start = tok;
while (Token::Match(start->previous(), "%type%|*|&"))
start = start->previous();
if (start != tok && start->isName() && (!start->previous() || Token::Match(start->previous(), "[;{}]")))
userFunctions.insert(tok->str());
}
if (userFunctions.find(tok->str()) == userFunctions.end() && stdFunctions.find(tok->str()) != stdFunctions.end())
insert = true;
} else if (Token::Match(tok, "%name% <") && stdTemplates.find(tok->str()) != stdTemplates.end())
insert = true;
else if (tok->isName() && !tok->varId() && !Token::Match(tok->next(), "(|<") && stdTypes.find(tok->str()) != stdTypes.end())
insert = true;
}
if (insert) {
tok->previous()->insertToken("std");
tok->previous()->linenr(tok->linenr()); // For stylistic reasons we put the std:: in the same line as the following token
tok->previous()->fileIndex(tok->fileIndex());
tok->previous()->insertToken("::");
} else if (isCPP11 && Token::Match(tok, "!!:: tr1 ::"))
tok->next()->str("std");
}
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (isCPP11 && Token::simpleMatch(tok, "std :: tr1 ::"))
Token::eraseTokens(tok, tok->tokAt(3));
else if (Token::simpleMatch(tok, "using namespace std ;")) {
Token::eraseTokens(tok, tok->tokAt(4));
tok->deleteThis();
}
}
}
void Tokenizer::simplifyMicrosoftMemoryFunctions()
{
// skip if not Windows
if (!mSettings->isWindowsPlatform())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->strAt(1) != "(")
continue;
if (Token::Match(tok, "CopyMemory|RtlCopyMemory|RtlCopyBytes")) {
tok->str("memcpy");
} else if (Token::Match(tok, "MoveMemory|RtlMoveMemory")) {
tok->str("memmove");
} else if (Token::Match(tok, "FillMemory|RtlFillMemory|RtlFillBytes")) {
// FillMemory(dst, len, val) -> memset(dst, val, len)
tok->str("memset");
Token *tok1 = tok->tokAt(2);
if (tok1)
tok1 = tok1->nextArgument(); // Second argument
if (tok1) {
Token *tok2 = tok1->nextArgument(); // Third argument
if (tok2)
Token::move(tok1->previous(), tok2->tokAt(-2), tok->next()->link()->previous()); // Swap third with second argument
}
} else if (Token::Match(tok, "ZeroMemory|RtlZeroMemory|RtlZeroBytes|RtlSecureZeroMemory")) {
// ZeroMemory(dst, len) -> memset(dst, 0, len)
tok->str("memset");
Token *tok1 = tok->tokAt(2);
if (tok1)
tok1 = tok1->nextArgument(); // Second argument
if (tok1) {
tok1 = tok1->previous();
tok1->insertToken("0");
tok1 = tok1->next();
tok1->insertToken(",");
}
} else if (Token::simpleMatch(tok, "RtlCompareMemory")) {
// RtlCompareMemory(src1, src2, len) -> memcmp(src1, src2, len)
tok->str("memcmp");
// For the record, when memcmp returns 0, both strings are equal.
// When RtlCompareMemory returns len, both strings are equal.
// It might be needed to improve this replacement by something
// like ((len - memcmp(src1, src2, len)) % (len + 1)) to
// respect execution path (if required)
}
}
}
namespace {
struct triplet {
triplet(const char* m, const char* u) : mbcs(m), unicode(u) {}
std::string mbcs, unicode;
};
const std::map<std::string, triplet> apis = {
std::make_pair("_topen", triplet("open", "_wopen")),
std::make_pair("_tsopen_s", triplet("_sopen_s", "_wsopen_s")),
std::make_pair("_tfopen", triplet("fopen", "_wfopen")),
std::make_pair("_tfopen_s", triplet("fopen_s", "_wfopen_s")),
std::make_pair("_tfreopen", triplet("freopen", "_wfreopen")),
std::make_pair("_tfreopen_s", triplet("freopen_s", "_wfreopen_s")),
std::make_pair("_tcscat", triplet("strcat", "wcscat")),
std::make_pair("_tcschr", triplet("strchr", "wcschr")),
std::make_pair("_tcscmp", triplet("strcmp", "wcscmp")),
std::make_pair("_tcsdup", triplet("strdup", "wcsdup")),
std::make_pair("_tcscpy", triplet("strcpy", "wcscpy")),
std::make_pair("_tcslen", triplet("strlen", "wcslen")),
std::make_pair("_tcsncat", triplet("strncat", "wcsncat")),
std::make_pair("_tcsncpy", triplet("strncpy", "wcsncpy")),
std::make_pair("_tcsnlen", triplet("strnlen", "wcsnlen")),
std::make_pair("_tcsrchr", triplet("strrchr", "wcsrchr")),
std::make_pair("_tcsstr", triplet("strstr", "wcsstr")),
std::make_pair("_tcstok", triplet("strtok", "wcstok")),
std::make_pair("_ftprintf", triplet("fprintf", "fwprintf")),
std::make_pair("_tprintf", triplet("printf", "wprintf")),
std::make_pair("_stprintf", triplet("sprintf", "swprintf")),
std::make_pair("_sntprintf", triplet("_snprintf", "_snwprintf")),
std::make_pair("_ftscanf", triplet("fscanf", "fwscanf")),
std::make_pair("_tscanf", triplet("scanf", "wscanf")),
std::make_pair("_stscanf", triplet("sscanf", "swscanf")),
std::make_pair("_ftprintf_s", triplet("fprintf_s", "fwprintf_s")),
std::make_pair("_tprintf_s", triplet("printf_s", "wprintf_s")),
std::make_pair("_stprintf_s", triplet("sprintf_s", "swprintf_s")),
std::make_pair("_sntprintf_s", triplet("_snprintf_s", "_snwprintf_s")),
std::make_pair("_ftscanf_s", triplet("fscanf_s", "fwscanf_s")),
std::make_pair("_tscanf_s", triplet("scanf_s", "wscanf_s")),
std::make_pair("_stscanf_s", triplet("sscanf_s", "swscanf_s"))
};
}
void Tokenizer::simplifyMicrosoftStringFunctions()
{
// skip if not Windows
if (!mSettings->isWindowsPlatform())
return;
const bool ansi = mSettings->platformType == Settings::Win32A;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->strAt(1) != "(")
continue;
const std::map<std::string, triplet>::const_iterator match = apis.find(tok->str());
if (match!=apis.end()) {
tok->str(ansi ? match->second.mbcs : match->second.unicode);
tok->originalName(match->first);
} else if (Token::Match(tok, "_T|_TEXT|TEXT ( %char%|%str% )")) {
tok->deleteNext();
tok->deleteThis();
tok->deleteNext();
if (!ansi)
tok->isLong(true);
while (Token::Match(tok->next(), "_T|_TEXT|TEXT ( %char%|%str% )")) {
tok->next()->deleteNext();
tok->next()->deleteThis();
tok->next()->deleteNext();
tok->concatStr(tok->next()->str());
tok->deleteNext();
}
}
}
}
// Remove Borland code
void Tokenizer::simplifyBorland()
{
// skip if not Windows
if (!mSettings->isWindowsPlatform())
return;
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "( __closure * %name% )")) {
tok->deleteNext();
}
}
// I think that these classes are always declared at the outer scope
// I save some time by ignoring inner classes.
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "{" && !Token::Match(tok->tokAt(-2), "namespace %type%")) {
tok = tok->link();
if (!tok)
break;
} else if (Token::Match(tok, "class %name% :|{")) {
while (tok && tok->str() != "{" && tok->str() != ";")
tok = tok->next();
if (!tok)
break;
if (tok->str() == ";")
continue;
const Token* end = tok->link()->next();
for (Token *tok2 = tok->next(); tok2 != end; tok2 = tok2->next()) {
if (tok2->str() == "__property" &&
Token::Match(tok2->previous(), ";|{|}|protected:|public:|__published:")) {
while (tok2->next() && !Token::Match(tok2->next(), "{|;"))
tok2->deleteNext();
tok2->deleteThis();
if (tok2->str() == "{") {
Token::eraseTokens(tok2, tok2->link());
tok2->deleteNext();
tok2->deleteThis();
// insert "; __property ;"
tok2->previous()->insertToken(";");
tok2->previous()->insertToken("__property");
tok2->previous()->insertToken(";");
}
}
}
}
}
}
// Remove Qt signals and slots
void Tokenizer::simplifyQtSignalsSlots()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
// check for emit which can be outside of class
if (Token::Match(tok, "emit|Q_EMIT %name% (") &&
Token::simpleMatch(tok->linkAt(2), ") ;")) {
tok->deleteThis();
} else if (!Token::Match(tok, "class %name% :"))
continue;
if (tok->previous() && tok->previous()->str() == "enum") {
tok = tok->tokAt(2);
continue;
}
// count { and } for tok2
unsigned int indentlevel = 0;
for (Token *tok2 = tok; tok2; tok2 = tok2->next()) {
if (tok2->str() == "{") {
++indentlevel;
if (indentlevel == 1)
tok = tok2;
else
tok2 = tok2->link();
} else if (tok2->str() == "}") {
if (indentlevel<2)
break;
else
--indentlevel;
}
if (tok2->strAt(1) == "Q_OBJECT")
tok2->deleteNext();
if (Token::Match(tok2->next(), "public|protected|private slots|Q_SLOTS :")) {
tok2 = tok2->next();
tok2->str(tok2->str() + ":");
tok2->deleteNext(2);
tok2 = tok2->previous();
} else if (Token::Match(tok2->next(), "signals|Q_SIGNALS :")) {
tok2 = tok2->next();
tok2->str("protected:");
tok2->deleteNext();
} else if (Token::Match(tok2->next(), "emit|Q_EMIT %name% (") &&
Token::simpleMatch(tok2->linkAt(3), ") ;")) {
tok2->deleteNext();
}
}
}
}
void Tokenizer::createSymbolDatabase()
{
if (!mSymbolDatabase)
mSymbolDatabase = new SymbolDatabase(this, mSettings, mErrorLogger);
mSymbolDatabase->validate();
}
void Tokenizer::deleteSymbolDatabase()
{
delete mSymbolDatabase;
mSymbolDatabase = nullptr;
}
static bool operatorEnd(const Token * tok)
{
if (tok && tok->str() == ")") {
tok = tok->next();
while (tok && !Token::Match(tok, "[=;{),]")) {
if (Token::Match(tok, "const|volatile")) {
tok = tok->next();
} else if (tok->str() == "noexcept") {
tok = tok->next();
if (tok && tok->str() == "(") {
tok = tok->link()->next();
}
} else if (tok->str() == "throw" && tok->next() && tok->next()->str() == "(") {
tok = tok->next()->link()->next();
}
// unknown macros ") MACRO {" and ") MACRO(...) {"
else if (tok->isUpperCaseName()) {
tok = tok->next();
if (tok && tok->str() == "(") {
tok = tok->link()->next();
}
} else
return false;
}
return true;
}
return false;
}
void Tokenizer::simplifyOperatorName()
{
if (isC())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "using|:: operator %op% ;")) {
tok->next()->str("operator" + tok->strAt(2));
tok->next()->deleteNext();
continue;
}
if (tok->str() != "operator")
continue;
// operator op
std::string op;
Token *par = tok->next();
bool done = false;
while (!done && par) {
done = true;
if (par->isName()) {
op += par->str();
par = par->next();
// merge namespaces eg. 'operator std :: string () const {'
if (Token::Match(par, ":: %name%|%op%|.")) {
op += par->str();
par = par->next();
}
done = false;
} else if (Token::Match(par, ".|%op%|,")) {
// check for operator in template
if (!(Token::Match(par, "<|>") && !op.empty())) {
op += par->str();
par = par->next();
done = false;
}
} else if (Token::simpleMatch(par, "[ ]")) {
op += "[]";
par = par->tokAt(2);
done = false;
} else if (Token::Match(par, "( *| )")) {
// break out and simplify..
if (operatorEnd(par->next()))
break;
while (par->str() != ")") {
op += par->str();
par = par->next();
}
op += ")";
par = par->next();
done = false;
}
}
if (par && (Token::Match(par, "<|>") || isFunctionHead(par, "{|;"))) {
tok->str("operator" + op);
Token::eraseTokens(tok, par);
}
if (!op.empty())
tok->isOperatorKeyword(true);
}
if (mSettings->debugwarnings) {
const Token *tok = list.front();
while ((tok = Token::findsimplematch(tok, "operator")) != nullptr) {
reportError(tok, Severity::debug, "debug",
"simplifyOperatorName: found unsimplified operator name");
tok = tok->next();
}
}
}
// remove unnecessary member qualification..
void Tokenizer::removeUnnecessaryQualification()
{
if (isC())
return;
std::vector<Space> classInfo;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "class|struct|namespace %type% :|{") &&
(!tok->previous() || tok->previous()->str() != "enum")) {
Space info;
info.isNamespace = tok->str() == "namespace";
tok = tok->next();
info.className = tok->str();
tok = tok->next();
while (tok && tok->str() != "{")
tok = tok->next();
if (!tok)
return;
info.bodyEnd = tok->link();
classInfo.push_back(info);
} else if (!classInfo.empty()) {
if (tok == classInfo.back().bodyEnd)
classInfo.pop_back();
else if (tok->str() == classInfo.back().className &&
!classInfo.back().isNamespace && tok->previous()->str() != ":" &&
(Token::Match(tok, "%type% :: ~| %type% (") ||
Token::Match(tok, "%type% :: operator"))) {
const Token *tok1 = tok->tokAt(3);
if (tok->strAt(2) == "operator") {
// check for operator ()
if (tok1->str() == "(")
tok1 = tok1->next();
while (tok1 && tok1->str() != "(") {
if (tok1->str() == ";")
break;
tok1 = tok1->next();
}
if (!tok1 || tok1->str() != "(")
continue;
} else if (tok->strAt(2) == "~")
tok1 = tok1->next();
if (!tok1 || !Token::Match(tok1->link(), ") const| {|;|:")) {
continue;
}
const bool isConstructorOrDestructor =
Token::Match(tok, "%type% :: ~| %type%") && (tok->strAt(2) == tok->str() || (tok->strAt(2) == "~" && tok->strAt(3) == tok->str()));
if (!isConstructorOrDestructor) {
bool isPrependedByType = Token::Match(tok->previous(), "%type%");
if (!isPrependedByType) {
const Token* tok2 = tok->tokAt(-2);
isPrependedByType = Token::Match(tok2, "%type% *|&");
}
if (!isPrependedByType) {
const Token* tok3 = tok->tokAt(-3);
isPrependedByType = Token::Match(tok3, "%type% * *|&");
}
if (!isPrependedByType) {
// It's not a constructor declaration and it's not a function declaration so
// this is a function call which can have all the qualifiers just fine - skip.
continue;
}
}
}
}
}
}
void Tokenizer::simplifyReturnStrncat()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::simpleMatch(tok, "return strncat (") &&
Token::simpleMatch(tok->linkAt(2), ") ;") &&
tok->strAt(3) != ")" && tok->strAt(3) != ",") {
//first argument
Token *tok2 = tok->tokAt(3);
//check if there are at least three arguments
for (unsigned char i = 0; i < 2; ++i) {
tok2 = tok2->nextArgument();
if (!tok2) {
tok = tok->linkAt(2)->next();
break;
}
}
if (!tok2)
continue;
tok2 = tok2->nextArgument();
//we want only three arguments
if (tok2) {
tok = tok->linkAt(2)->next();
continue;
}
// Remove 'return'
tok->deleteThis();
// Add 'return arg1 ;' after 'strncat(arg1, arg2, arg3);'
tok = tok->next();
tok2 = tok->link()->next();
tok2->insertToken(";");
//the last token of the first argument before ','
const Token * const end = tok->next()->nextArgument()->tokAt(-2);
//all the first argument is copied
TokenList::copyTokens(tok2, tok->next(), end);
tok2->insertToken("return");
}
}
}
void Tokenizer::printUnknownTypes() const
{
if (!mSymbolDatabase)
return;
std::multimap<std::string, const Token *> unknowns;
for (unsigned int i = 1; i <= mVarId; ++i) {
const Variable *var = mSymbolDatabase->getVariableFromVarId(i);
if (!var)
continue;
// is unknown type?
if (var->type() || var->typeStartToken()->isStandardType())
continue;
std::string name;
const Token * nameTok;
// single token type?
if (var->typeStartToken() == var->typeEndToken()) {
nameTok = var->typeStartToken();
name = nameTok->str();
}
// complicated type
else {
const Token *tok = var->typeStartToken();
int level = 0;
nameTok = tok;
while (tok) {
// skip pointer and reference part of type
if (level == 0 && Token::Match(tok, "*|&"))
break;
name += tok->str();
if (Token::Match(tok, "struct|union|enum"))
name += " ";
// pointers and references are OK in template
else if (tok->str() == "<")
++level;
else if (tok->str() == ">")
--level;
if (tok == var->typeEndToken())
break;
tok = tok->next();
}
}
unknowns.insert(std::pair<std::string, const Token *>(name, nameTok));
}
if (!unknowns.empty()) {
std::string last;
size_t count = 0;
for (std::multimap<std::string, const Token *>::const_iterator it = unknowns.begin(); it != unknowns.end(); ++it) {
// skip types is std namespace because they are not interesting
if (it->first.find("std::") != 0) {
if (it->first != last) {
last = it->first;
count = 1;
reportError(it->second, Severity::debug, "debug", "Unknown type \'" + it->first + "\'.");
} else {
if (count < 3) // limit same type to 3
reportError(it->second, Severity::debug, "debug", "Unknown type \'" + it->first + "\'.");
count++;
}
}
}
}
}
void Tokenizer::simplifyMathExpressions()
{
for (Token *tok = list.front(); tok; tok = tok->next()) {
//simplify Pythagorean trigonometric identity: pow(sin(x),2)+pow(cos(x),2) = 1
// pow(cos(x),2)+pow(sin(x),2) = 1
// @todo: sin(x) * sin(x) + cos(x) * cos(x) = 1
// cos(x) * cos(x) + sin(x) * sin(x) = 1
//simplify Hyperbolic identity: pow(sinh(x),2)-pow(cosh(x),2) = -1
// pow(cosh(x),2)-pow(sinh(x),2) = -1
// @todo: sinh(x) * sinh(x) - cosh(x) * cosh(x) = -1
// cosh(x) * cosh(x) - sinh(x) * sinh(x) = -1
if (Token::Match(tok, "pow|powf|powl (")) {
if (Token::Match(tok->tokAt(2), "sin|sinf|sinl (")) {
Token * const tok2 = tok->linkAt(3);
if (!Token::Match(tok2, ") , %num% ) + pow|powf|powl ( cos|cosf|cosl ("))
continue;
const std::string& leftExponent = tok2->strAt(2);
if (!isTwoNumber(leftExponent))
continue; // left exponent is not 2
const Token * const tok3 = tok2->tokAt(8);
Token * const tok4 = tok3->link();
if (!Token::Match(tok4, ") , %num% )"))
continue;
const std::string& rightExponent = tok4->strAt(2);
if (!isTwoNumber(rightExponent))
continue; // right exponent is not 2
if (tok->tokAt(3)->stringifyList(tok2->next()) == tok3->stringifyList(tok4->next())) {
Token::eraseTokens(tok, tok4->tokAt(4));
tok->str("1");
}
} else if (Token::Match(tok->tokAt(2), "cos|cosf|cosl (")) {
Token * const tok2 = tok->linkAt(3);
if (!Token::Match(tok2, ") , %num% ) + pow|powf|powl ( sin|sinf|sinl ("))
continue;
const std::string& leftExponent = tok2->strAt(2);
if (!isTwoNumber(leftExponent))
continue; // left exponent is not 2
const Token * const tok3 = tok2->tokAt(8);
Token * const tok4 = tok3->link();
if (!Token::Match(tok4, ") , %num% )"))
continue;
const std::string& rightExponent = tok4->strAt(2);
if (!isTwoNumber(rightExponent))
continue; // right exponent is not 2
if (tok->tokAt(3)->stringifyList(tok2->next()) == tok3->stringifyList(tok4->next())) {
Token::eraseTokens(tok, tok4->tokAt(4));
tok->str("1");
}
} else if (Token::Match(tok->tokAt(2), "sinh|sinhf|sinhl (")) {
Token * const tok2 = tok->linkAt(3);
if (!Token::Match(tok2, ") , %num% ) - pow|powf|powl ( cosh|coshf|coshl ("))
continue;
const std::string& leftExponent = tok2->strAt(2);
if (!isTwoNumber(leftExponent))
continue; // left exponent is not 2
const Token * const tok3 = tok2->tokAt(8);
Token * const tok4 = tok3->link();
if (!Token::Match(tok4, ") , %num% )"))
continue;
const std::string& rightExponent = tok4->strAt(2);
if (!isTwoNumber(rightExponent))
continue; // right exponent is not 2
if (tok->tokAt(3)->stringifyList(tok2->next()) == tok3->stringifyList(tok4->next())) {
Token::eraseTokens(tok, tok4->tokAt(4));
tok->str("-1");
}
} else if (Token::Match(tok->tokAt(2), "cosh|coshf|coshl (")) {
Token * const tok2 = tok->linkAt(3);
if (!Token::Match(tok2, ") , %num% ) - pow|powf|powl ( sinh|sinhf|sinhl ("))
continue;
const std::string& leftExponent = tok2->strAt(2);
if (!isTwoNumber(leftExponent))
continue; // left exponent is not 2
const Token * const tok3 = tok2->tokAt(8);
Token * const tok4 = tok3->link();
if (!Token::Match(tok4, ") , %num% )"))
continue;
const std::string& rightExponent = tok4->strAt(2);
if (!isTwoNumber(rightExponent))
continue; // right exponent is not 2
if (tok->tokAt(3)->stringifyList(tok2->next()) == tok3->stringifyList(tok4->next())) {
Token::eraseTokens(tok, tok4->tokAt(4));
tok->str("-1");
}
}
}
}
}
bool Tokenizer::simplifyStrlen()
{
// replace strlen(str)
bool modified=false;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "strlen ( %str% )")) {
tok->str(MathLib::toString(Token::getStrLength(tok->tokAt(2))));
tok->deleteNext(3);
modified=true;
}
}
return modified;
}
void Tokenizer::prepareTernaryOpForAST()
{
// http://en.cppreference.com/w/cpp/language/operator_precedence says about ternary operator:
// "The expression in the middle of the conditional operator (between ? and :) is parsed as if parenthesized: its precedence relative to ?: is ignored."
// The AST parser relies on this function to add such parentheses where necessary.
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "?") {
bool parenthesesNeeded = false;
unsigned int depth = 0;
Token* tok2 = tok->next();
for (; tok2; tok2 = tok2->next()) {
if (tok2->link() && Token::Match(tok2, "[|(|<"))
tok2 = tok2->link();
else if (tok2->str() == ":") {
if (depth == 0)
break;
depth--;
} else if (tok2->str() == ";" || (tok2->link() && tok2->str() != "{" && tok2->str() != "}"))
break;
else if (tok2->str() == ",")
parenthesesNeeded = true;
else if (tok2->str() == "<")
parenthesesNeeded = true;
else if (tok2->str() == "?") {
depth++;
parenthesesNeeded = true;
}
}
if (parenthesesNeeded && tok2 && tok2->str() == ":") {
tok->insertToken("(");
tok2->insertToken(")", emptyString, true);
Token::createMutualLinks(tok->next(), tok2->previous());
}
}
}
}
void Tokenizer::reportError(const Token* tok, const Severity::SeverityType severity, const std::string& id, const std::string& msg, bool inconclusive) const
{
const std::list<const Token*> callstack(1, tok);
reportError(callstack, severity, id, msg, inconclusive);
}
void Tokenizer::reportError(const std::list<const Token*>& callstack, Severity::SeverityType severity, const std::string& id, const std::string& msg, bool inconclusive) const
{
const ErrorLogger::ErrorMessage errmsg(callstack, &list, severity, id, msg, inconclusive);
if (mErrorLogger)
mErrorLogger->reportErr(errmsg);
else
Check::reportError(errmsg);
}
void Tokenizer::setPodTypes()
{
if (!mSettings)
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (!tok->isName())
continue;
// pod type
const struct Library::PodType *podType = mSettings->library.podtype(tok->str());
if (podType) {
const Token *prev = tok->previous();
while (prev && prev->isName())
prev = prev->previous();
if (prev && !Token::Match(prev, ";|{|}|,|("))
continue;
tok->isStandardType(true);
}
}
}
void Tokenizer::SimplifyNamelessRValueReferences()
{
// Simplify nameless rValue references - named ones are simplified later
for (Token* tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "&& [,)]")) {
tok->str("&");
tok->insertToken("&");
}
}
}
const Token *Tokenizer::findSQLBlockEnd(const Token *tokSQLStart)
{
const Token *tokLastEnd = nullptr;
for (const Token *tok = tokSQLStart->tokAt(2); tok != nullptr; tok = tok->next()) {
if (tokLastEnd == nullptr && tok->str() == ";")
tokLastEnd = tok;
else if (tok->str() == "__CPPCHECK_EMBEDDED_SQL_EXEC__") {
if (Token::simpleMatch(tok->tokAt(-2), "END - __CPPCHECK_EMBEDDED_SQL_EXEC__ ;"))
return tok->next();
return tokLastEnd;
} else if (Token::Match(tok, "{|}|==|&&|!|^|<<|>>|++|+=|-=|/=|*=|>>=|<<=|~"))
break; // We are obviously outside the SQL block
}
return tokLastEnd;
}
void Tokenizer::simplifyNestedNamespace()
{
if (!isCPP())
return;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (Token::Match(tok, "namespace %name% ::") && tok->strAt(-1) != "using") {
Token * tok2 = tok->tokAt(2);
// validate syntax
while (Token::Match(tok2, ":: %name%"))
tok2 = tok2->tokAt(2);
if (!tok2 || tok2->str() != "{")
return; // syntax error
std::stack<Token *> links;
tok2 = tok->tokAt(2);
while (tok2->str() == "::") {
links.push(tok2);
tok2->str("{");
tok2->insertToken("namespace");
tok2 = tok2->tokAt(3);
}
tok = tok2;
if (!links.empty() && tok2->str() == "{") {
tok2 = tok2->link();
while (!links.empty()) {
tok2->insertToken("}");
tok2 = tok2->next();
Token::createMutualLinks(links.top(), tok2);
links.pop();
}
}
}
}
}
static bool sameTokens(const Token *first, const Token *last, const Token *other)
{
while (other && first->str() == other->str()) {
if (first == last)
return true;
first = first->next();
other = other->next();
}
return false;
}
static Token * deleteAlias(Token * tok)
{
Token::eraseTokens(tok, Token::findsimplematch(tok, ";"));
// delete first token
tok->deleteThis();
// delete ';' if not last token
tok->deleteThis();
return tok;
}
void Tokenizer::simplifyNamespaceAliases()
{
if (!isCPP())
return;
int scope = 0;
for (Token *tok = list.front(); tok; tok = tok->next()) {
if (tok->str() == "{")
scope++;
else if (tok->str() == "}")
scope--;
else if (Token::Match(tok, "namespace %name% =")) {
const std::string name(tok->next()->str());
Token * tokNameStart = tok->tokAt(3);
Token * tokNameEnd = tokNameStart;
while (tokNameEnd && tokNameEnd->next() && tokNameEnd->next()->str() != ";")
tokNameEnd = tokNameEnd->next();
if (!tokNameEnd)
return; // syntax error
int endScope = scope;
Token * tokLast = tokNameEnd->next();
Token * tokNext = tokLast->next();
Token * tok2 = tokNext;
while (tok2 && endScope >= scope) {
if (Token::simpleMatch(tok2, "{"))
endScope++;
else if (Token::simpleMatch(tok2, "}"))
endScope--;
else if (tok2->str() == name) {
if (Token::Match(tok2->previous(), "namespace %name% =")) {
// check for possible duplicate aliases
if (sameTokens(tokNameStart, tokNameEnd, tok2->tokAt(2))) {
// delete duplicate
tok2 = deleteAlias(tok2->previous());
continue;
} else {
// conflicting declaration (syntax error)
if (endScope == scope) {
// delete conflicting declaration
tok2 = deleteAlias(tok2->previous());
}
// new declaration
else {
// TODO: use the new alias in this scope
tok2 = deleteAlias(tok2->previous());
}
continue;
}
}
tok2->str(tokNameStart->str());
Token * tok3 = tokNameStart;
while (tok3 != tokNameEnd) {
tok2->insertToken(tok3->next()->str());
tok2 = tok2->next();
tok3 = tok3->next();
}
}
tok2 = tok2->next();
}
if (tok->previous() && tokNext) {
Token::eraseTokens(tok->previous(), tokNext);
tok = tokNext->previous();
} else if (tok->previous()) {
Token::eraseTokens(tok->previous(), tokLast);
tok = tokLast;
} else if (tokNext) {
Token::eraseTokens(tok, tokNext);
tok->deleteThis();
} else {
Token::eraseTokens(tok, tokLast);
tok->deleteThis();
}
}
}
}
Tokenizer::VariableMap::VariableMap() : mVarId(0) {}
void Tokenizer::VariableMap::enterScope()
{
mScopeInfo.push(std::list<std::pair<std::string, unsigned int>>());
}
bool Tokenizer::VariableMap::leaveScope()
{
if (mScopeInfo.empty())
return false;
for (const std::pair<std::string, unsigned int> &outerVariable : mScopeInfo.top()) {
if (outerVariable.second != 0)
mVariableId[outerVariable.first] = outerVariable.second;
else
mVariableId.erase(outerVariable.first);
}
mScopeInfo.pop();
return true;
}
void Tokenizer::VariableMap::addVariable(const std::string &varname)
{
if (mScopeInfo.empty()) {
mVariableId[varname] = ++mVarId;
return;
}
std::map<std::string, unsigned int>::iterator it = mVariableId.find(varname);
if (it == mVariableId.end()) {
mScopeInfo.top().push_back(std::pair<std::string, unsigned int>(varname, 0));
mVariableId[varname] = ++mVarId;
return;
}
mScopeInfo.top().push_back(std::pair<std::string, unsigned int>(varname, it->second));
it->second = ++mVarId;
}
bool Tokenizer::VariableMap::hasVariable(const std::string &varname) const
{
return mVariableId.find(varname) != mVariableId.end();
}