cppcheck/lib/tokenlist.cpp

1324 lines
46 KiB
C++
Raw Normal View History

/*
* Cppcheck - A tool for static C/C++ code analysis
2018-01-14 15:37:52 +01:00
* Copyright (C) 2007-2018 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 "tokenlist.h"
2017-05-27 04:33:47 +02:00
#include "errorlogger.h"
#include "mathlib.h"
#include "path.h"
#include "settings.h"
2017-05-27 04:33:47 +02:00
#include "token.h"
2017-05-18 22:25:49 +02:00
#include <simplecpp.h>
#include <cctype>
2017-05-27 04:33:47 +02:00
#include <cstring>
#include <set>
#include <stack>
// How many compileExpression recursions are allowed?
// For practical code this could be endless. But in some special torture test
// there needs to be a limit.
static const unsigned int AST_MAX_DEPTH = 50U;
TokenList::TokenList(const Settings* settings) :
mTokensFrontBack(),
mSettings(settings),
2018-06-16 16:28:03 +02:00
mIsC(false),
mIsCpp(false)
{
}
TokenList::~TokenList()
{
deallocateTokens();
}
//---------------------------------------------------------------------------
const std::string& TokenList::getSourceFilePath() const
{
if (getFiles().empty()) {
return emptyString;
}
return getFiles()[0];
}
//---------------------------------------------------------------------------
// Deallocate lists..
void TokenList::deallocateTokens()
{
deleteTokens(mTokensFrontBack.front);
mTokensFrontBack.front = nullptr;
mTokensFrontBack.back = nullptr;
2018-06-16 16:23:55 +02:00
mFiles.clear();
}
unsigned int TokenList::appendFileIfNew(const std::string &fileName)
{
// Has this file been tokenized already?
2018-06-16 16:23:55 +02:00
for (std::size_t i = 0; i < mFiles.size(); ++i)
if (Path::sameFileName(mFiles[i], fileName))
return (unsigned int)i;
2018-06-16 16:23:55 +02:00
// The "mFiles" vector remembers what files have been tokenized..
mFiles.push_back(Path::simplifyPath(fileName));
2018-06-16 16:28:03 +02:00
// Update mIsC and mIsCpp properties
2018-06-16 16:23:55 +02:00
if (mFiles.size() == 1) { // Update only useful if first file added to _files
if (!mSettings) {
2018-06-16 16:28:03 +02:00
mIsC = Path::isC(getSourceFilePath());
mIsCpp = Path::isCPP(getSourceFilePath());
} else {
2018-06-16 16:28:03 +02:00
mIsC = mSettings->enforcedLang == Settings::C || (mSettings->enforcedLang == Settings::None && Path::isC(getSourceFilePath()));
mIsCpp = mSettings->enforcedLang == Settings::CPP || (mSettings->enforcedLang == Settings::None && Path::isCPP(getSourceFilePath()));
}
}
2018-06-16 16:23:55 +02:00
return mFiles.size() - 1U;
}
void TokenList::deleteTokens(Token *tok)
{
while (tok) {
Token *next = tok->next();
delete tok;
tok = next;
}
}
//---------------------------------------------------------------------------
// add a token.
//---------------------------------------------------------------------------
void TokenList::addtoken(std::string str, const unsigned int lineno, const unsigned int fileno, bool split)
{
2014-03-09 17:54:49 +01:00
if (str.empty())
return;
// If token contains # characters, split it up
2014-03-09 17:54:49 +01:00
if (split) {
size_t begin = 0;
size_t end = 0;
while ((end = str.find("##", begin)) != std::string::npos) {
addtoken(str.substr(begin, end - begin), lineno, fileno, false);
addtoken("##", lineno, fileno, false);
begin = end+2;
}
if (begin != 0) {
addtoken(str.substr(begin), lineno, fileno, false);
return;
}
}
// Replace hexadecimal value with decimal
2017-05-06 11:57:02 +02:00
const bool isHex = MathLib::isIntHex(str) ;
if (isHex || MathLib::isOct(str) || MathLib::isBin(str)) {
// TODO: It would be better if TokenList didn't simplify hexadecimal numbers
std::string suffix;
if (isHex &&
str.size() == (2 + mSettings->int_bit / 4) &&
(str[2] >= '8') && // includes A-F and a-f
2017-05-06 11:57:02 +02:00
MathLib::getSuffix(str).empty()
)
suffix = "U";
str = MathLib::value(str).str() + suffix;
}
if (mTokensFrontBack.back) {
mTokensFrontBack.back->insertToken(str);
} else {
mTokensFrontBack.front = new Token(&mTokensFrontBack);
mTokensFrontBack.back = mTokensFrontBack.front;
mTokensFrontBack.back->str(str);
}
if (isCPP() && str == "delete")
mTokensFrontBack.back->isKeyword(true);
mTokensFrontBack.back->linenr(lineno);
mTokensFrontBack.back->fileIndex(fileno);
}
void TokenList::addtoken(const Token * tok, const unsigned int lineno, const unsigned int fileno)
{
if (tok == nullptr)
return;
if (mTokensFrontBack.back) {
mTokensFrontBack.back->insertToken(tok->str(), tok->originalName());
} else {
mTokensFrontBack.front = new Token(&mTokensFrontBack);
mTokensFrontBack.back = mTokensFrontBack.front;
mTokensFrontBack.back->str(tok->str());
if (!tok->originalName().empty())
mTokensFrontBack.back->originalName(tok->originalName());
}
mTokensFrontBack.back->linenr(lineno);
mTokensFrontBack.back->fileIndex(fileno);
mTokensFrontBack.back->flags(tok->flags());
}
//---------------------------------------------------------------------------
// copyTokens - Copy and insert tokens
//---------------------------------------------------------------------------
Token *TokenList::copyTokens(Token *dest, const Token *first, const Token *last, bool one_line)
{
std::stack<Token *> links;
Token *tok2 = dest;
unsigned int linenrs = dest->linenr();
const unsigned int commonFileIndex = dest->fileIndex();
for (const Token *tok = first; tok != last->next(); tok = tok->next()) {
tok2->insertToken(tok->str());
tok2 = tok2->next();
tok2->fileIndex(commonFileIndex);
tok2->linenr(linenrs);
tok2->tokType(tok->tokType());
tok2->flags(tok->flags());
tok2->varId(tok->varId());
// Check for links and fix them up
if (Token::Match(tok2, "(|[|{"))
links.push(tok2);
else if (Token::Match(tok2, ")|]|}")) {
if (links.empty())
return tok2;
Token * link = links.top();
tok2->link(link);
link->link(tok2);
links.pop();
}
if (!one_line && tok->next())
linenrs += tok->next()->linenr() - tok->linenr();
}
return tok2;
}
//---------------------------------------------------------------------------
// InsertTokens - Copy and insert tokens
//---------------------------------------------------------------------------
void TokenList::insertTokens(Token *dest, const Token *src, unsigned int n)
{
std::stack<Token *> link;
while (n > 0) {
dest->insertToken(src->str(), src->originalName());
dest = dest->next();
// Set links
if (Token::Match(dest, "(|[|{"))
link.push(dest);
else if (!link.empty() && Token::Match(dest, ")|]|}")) {
Token::createMutualLinks(dest, link.top());
link.pop();
}
dest->fileIndex(src->fileIndex());
dest->linenr(src->linenr());
dest->varId(src->varId());
dest->tokType(src->tokType());
dest->flags(src->flags());
src = src->next();
--n;
}
}
//---------------------------------------------------------------------------
// Tokenize - tokenizes a given file.
//---------------------------------------------------------------------------
bool TokenList::createTokens(std::istream &code, const std::string& file0)
{
appendFileIfNew(file0);
2017-05-18 21:52:31 +02:00
simplecpp::OutputList outputList;
2018-06-16 16:23:55 +02:00
simplecpp::TokenList tokens(code, mFiles, file0, &outputList);
2017-05-18 21:52:31 +02:00
createTokens(&tokens);
2017-05-18 21:52:31 +02:00
return outputList.empty();
}
//---------------------------------------------------------------------------
void TokenList::createTokens(const simplecpp::TokenList *tokenList)
{
if (tokenList->cfront())
2018-06-16 16:23:55 +02:00
mFiles = tokenList->cfront()->location.files;
else
2018-06-16 16:23:55 +02:00
mFiles.clear();
2018-06-16 16:28:03 +02:00
mIsC = mIsCpp = false;
2018-06-16 16:23:55 +02:00
if (!mFiles.empty()) {
2018-06-16 16:28:03 +02:00
mIsC = Path::isC(getSourceFilePath());
mIsCpp = Path::isCPP(getSourceFilePath());
}
if (mSettings && mSettings->enforcedLang != Settings::None) {
2018-06-16 16:28:03 +02:00
mIsC = (mSettings->enforcedLang == Settings::C);
mIsCpp = (mSettings->enforcedLang == Settings::CPP);
}
for (const simplecpp::Token *tok = tokenList->cfront(); tok; tok = tok->next) {
2017-05-18 21:52:31 +02:00
2018-05-14 13:00:22 +02:00
std::string str = tok->str();
2017-05-18 21:52:31 +02:00
// Replace hexadecimal value with decimal
// TODO: Remove this
const bool isHex = MathLib::isIntHex(str) ;
if (isHex || MathLib::isOct(str) || MathLib::isBin(str)) {
// TODO: It would be better if TokenList didn't simplify hexadecimal numbers
std::string suffix;
if (isHex &&
mSettings &&
str.size() == (2 + mSettings->int_bit / 4) &&
2017-05-18 21:52:31 +02:00
(str[2] >= '8') && // includes A-F and a-f
MathLib::getSuffix(str).empty()
)
suffix = "U";
str = MathLib::value(str).str() + suffix;
}
// Float literal
if (str.size() > 1 && str[0] == '.' && std::isdigit(str[1]))
str = '0' + str;
if (mTokensFrontBack.back) {
mTokensFrontBack.back->insertToken(str);
} else {
mTokensFrontBack.front = new Token(&mTokensFrontBack);
mTokensFrontBack.back = mTokensFrontBack.front;
mTokensFrontBack.back->str(str);
}
if (isCPP() && mTokensFrontBack.back->str() == "delete")
mTokensFrontBack.back->isKeyword(true);
mTokensFrontBack.back->fileIndex(tok->location.fileIndex);
mTokensFrontBack.back->linenr(tok->location.line);
mTokensFrontBack.back->col(tok->location.col);
mTokensFrontBack.back->isExpandedMacro(!tok->macro.empty());
}
if (mSettings && mSettings->relativePaths) {
2018-06-16 16:23:55 +02:00
for (std::size_t i = 0; i < mFiles.size(); i++)
mFiles[i] = Path::getRelativePath(mFiles[i], mSettings->basePaths);
}
Token::assignProgressValues(mTokensFrontBack.front);
}
//---------------------------------------------------------------------------
unsigned long long TokenList::calculateChecksum() const
{
unsigned long long checksum = 0;
for (const Token* tok = front(); tok; tok = tok->next()) {
const unsigned int subchecksum1 = tok->flags() + tok->varId() + tok->tokType();
unsigned int subchecksum2 = 0;
for (std::size_t i = 0; i < tok->str().size(); i++)
subchecksum2 += (unsigned int)tok->str()[i];
if (!tok->originalName().empty()) {
for (std::size_t i = 0; i < tok->originalName().size(); i++)
subchecksum2 += (unsigned int) tok->originalName()[i];
}
checksum ^= ((static_cast<unsigned long long>(subchecksum1) << 32) | subchecksum2);
const bool bit1 = (checksum & 1) != 0;
checksum >>= 1;
if (bit1)
checksum |= (1ULL << 63);
}
return checksum;
}
//---------------------------------------------------------------------------
struct AST_state {
std::stack<Token*> op;
unsigned int depth;
unsigned int inArrayAssignment;
bool cpp;
unsigned int assign;
2017-06-08 15:32:35 +02:00
bool inCase;
explicit AST_state(bool cpp_) : depth(0), inArrayAssignment(0), cpp(cpp_), assign(0U), inCase(false) {}
};
static Token * skipDecl(Token *tok)
{
if (!Token::Match(tok->previous(), "( %name%"))
return tok;
Token *vartok = tok;
while (Token::Match(vartok, "%name%|*|&|::|<")) {
if (vartok->str() == "<") {
if (vartok->link())
vartok = vartok->link();
else
return tok;
} else if (Token::Match(vartok, "%name% [:=]")) {
return vartok;
}
vartok = vartok->next();
}
return tok;
}
2013-11-11 16:39:34 +01:00
static bool iscast(const Token *tok)
{
if (!Token::Match(tok, "( ::| %name%"))
2013-11-11 16:39:34 +01:00
return false;
if (Token::simpleMatch(tok->link(), ") ( )"))
return false;
if (tok->previous() && tok->previous()->isName() && tok->previous()->str() != "return")
return false;
if (Token::simpleMatch(tok->previous(), ">") && tok->previous()->link())
2016-01-03 09:38:03 +01:00
return false;
if (Token::Match(tok, "( (| typeof (") && Token::Match(tok->link(), ") %num%"))
return true;
if (Token::Match(tok->link(), ") }|)|]"))
return false;
if (Token::Match(tok->link(), ") %cop%") && !Token::Match(tok->link(), ") [&*+-~]"))
return false;
if (Token::Match(tok->previous(), "= ( %name% ) {") && tok->next()->varId() == 0)
return true;
bool type = false;
2013-11-11 16:39:34 +01:00
for (const Token *tok2 = tok->next(); tok2; tok2 = tok2->next()) {
2014-05-24 20:21:08 +02:00
while (tok2->link() && Token::Match(tok2, "(|[|<"))
tok2 = tok2->link()->next();
if (tok2->str() == ")") {
return type || tok2->strAt(-1) == "*" || Token::simpleMatch(tok2, ") ~") ||
(Token::Match(tok2, ") %any%") &&
2015-10-17 18:25:27 +02:00
!tok2->next()->isOp() &&
!Token::Match(tok2->next(), "[[]);,?:.]"));
}
if (!Token::Match(tok2, "%name%|*|&|::"))
return false;
2017-04-10 23:44:30 +02:00
if (tok2->isStandardType() && (tok2->next()->str() != "(" || Token::Match(tok2->next(), "( * *| )")))
type = true;
2013-11-11 16:39:34 +01:00
}
return false;
}
// int(1), int*(2), ..
static Token * findCppTypeInitPar(Token *tok)
{
if (!tok || !Token::Match(tok->previous(), "[,()] %name%"))
return nullptr;
bool istype = false;
while (Token::Match(tok, "%name%|::|<")) {
if (tok->str() == "<") {
tok = tok->link();
if (!tok)
return nullptr;
}
istype |= tok->isStandardType();
tok = tok->next();
}
if (!istype)
return nullptr;
if (!Token::Match(tok, "[*&]"))
return nullptr;
while (Token::Match(tok, "[*&]"))
tok = tok->next();
return (tok && tok->str() == "(") ? tok : nullptr;
}
// X{} X<Y>{} etc
static bool iscpp11init(const Token * const tok)
{
const Token *nameToken = tok;
while (nameToken && nameToken->str() == "{") {
nameToken = nameToken->previous();
if (nameToken && nameToken->str() == "," && Token::simpleMatch(nameToken->previous(), "} ,"))
nameToken = nameToken->linkAt(-1);
}
if (!nameToken)
return false;
if (nameToken->str() == ">" && nameToken->link())
nameToken = nameToken->link()->previous();
2017-04-10 23:05:41 +02:00
const Token *endtok = nullptr;
if (Token::Match(nameToken, "%name% { !!["))
endtok = nameToken->linkAt(1);
else if (Token::Match(nameToken,"%name% <") && Token::simpleMatch(nameToken->linkAt(1),"> {"))
endtok = nameToken->linkAt(1)->linkAt(1);
else
return false;
// There is no initialisation for example here: 'class Fred {};'
if (!Token::simpleMatch(endtok, "} ;"))
return true;
const Token *prev = nameToken;
while (Token::Match(prev, "%name%|::|:|<|>")) {
if (Token::Match(prev, "class|struct"))
return false;
prev = prev->previous();
}
return true;
}
static void compileUnaryOp(Token *&tok, AST_state& state, void(*f)(Token *&tok, AST_state& state))
{
2013-11-04 11:26:16 +01:00
Token *unaryop = tok;
if (f) {
tok = tok->next();
state.depth++;
if (tok && state.depth <= AST_MAX_DEPTH)
f(tok, state);
state.depth--;
}
2013-11-04 11:26:16 +01:00
if (!state.op.empty()) {
unaryop->astOperand1(state.op.top());
state.op.pop();
2013-11-04 11:26:16 +01:00
}
state.op.push(unaryop);
2013-11-04 11:26:16 +01:00
}
static void compileBinOp(Token *&tok, AST_state& state, void(*f)(Token *&tok, AST_state& state))
2013-11-04 11:26:16 +01:00
{
Token *binop = tok;
if (f) {
tok = tok->next();
state.depth++;
if (tok && state.depth <= AST_MAX_DEPTH)
f(tok, state);
state.depth--;
}
2013-11-04 11:26:16 +01:00
// TODO: Should we check if op is empty.
// * Is it better to add assertion that it isn't?
// * Write debug warning if it's empty?
if (!state.op.empty()) {
binop->astOperand2(state.op.top());
state.op.pop();
2013-11-04 11:26:16 +01:00
}
if (!state.op.empty()) {
binop->astOperand1(state.op.top());
state.op.pop();
2013-11-04 11:26:16 +01:00
}
state.op.push(binop);
2013-11-04 11:26:16 +01:00
}
static void compileExpression(Token *&tok, AST_state& state);
2013-11-04 11:26:16 +01:00
static void compileTerm(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
2013-11-11 16:39:34 +01:00
if (!tok)
return;
2013-11-20 05:57:56 +01:00
if (Token::Match(tok, "L %str%|%char%"))
tok = tok->next();
2017-04-21 21:02:46 +02:00
if (state.inArrayAssignment && Token::Match(tok->previous(), "[{,] . %name%")) { // Jump over . in C style struct initialization
state.op.push(tok);
tok->astOperand1(tok->next());
tok = tok->tokAt(2);
}
if (state.inArrayAssignment && Token::Match(tok->previous(), "[{,] [ %num%|%name% ]")) {
state.op.push(tok);
tok->astOperand1(tok->next());
tok = tok->tokAt(3);
}
2013-11-04 11:26:16 +01:00
if (tok->isLiteral()) {
state.op.push(tok);
do {
tok = tok->next();
} while (Token::Match(tok, "%name%|%str%"));
2017-06-08 15:32:35 +02:00
} else if (tok->isName()) {
if (Token::Match(tok, "return|case") || (state.cpp && tok->str() == "throw")) {
2017-06-08 15:32:35 +02:00
if (tok->str() == "case")
state.inCase = true;
compileUnaryOp(tok, state, compileExpression);
state.op.pop();
2017-06-08 15:32:35 +02:00
if (state.inCase && Token::simpleMatch(tok, ": ;"))
tok = tok->next();
} else if (Token::Match(tok, "sizeof !!(")) {
compileUnaryOp(tok, state, compileExpression);
state.op.pop();
} else if (state.cpp && findCppTypeInitPar(tok)) { // int(0), int*(123), ..
tok = findCppTypeInitPar(tok);
state.op.push(tok);
tok = tok->tokAt(2);
} else if (state.cpp && iscpp11init(tok)) { // X{} X<Y>{} etc
state.op.push(tok);
tok = tok->next();
if (tok->str() == "<")
tok = tok->link()->next();
} else if (!state.cpp || !Token::Match(tok, "new|delete %name%|*|&|::|(|[")) {
tok = skipDecl(tok);
while (tok->next() && tok->next()->isName())
tok = tok->next();
state.op.push(tok);
if (Token::Match(tok, "%name% <") && tok->linkAt(1))
tok = tok->linkAt(1);
2013-11-04 11:26:16 +01:00
tok = tok->next();
if (Token::Match(tok, "%str%")) {
while (Token::Match(tok, "%name%|%str%"))
tok = tok->next();
}
}
2014-02-05 06:05:48 +01:00
} else if (tok->str() == "{") {
const Token *prev = tok->previous();
2017-04-21 21:02:46 +02:00
if (Token::simpleMatch(prev, ") {") && iscast(prev->link()))
prev = prev->link()->previous();
2017-04-10 22:17:34 +02:00
if (Token::simpleMatch(tok->link(),"} [")) {
tok = tok->next();
} else if (state.cpp && iscpp11init(tok)) {
if (state.op.empty() || Token::Match(tok->previous(), "[{,]"))
compileUnaryOp(tok, state, compileExpression);
else
compileBinOp(tok, state, compileExpression);
if (Token::Match(tok, "} ,|:")) {
tok = tok->next();
}
2017-04-21 21:02:46 +02:00
} else if (!state.inArrayAssignment && !Token::simpleMatch(prev, "=")) {
state.op.push(tok);
tok = tok->link()->next();
} else {
if (tok->link() != tok->next()) {
state.inArrayAssignment++;
compileUnaryOp(tok, state, compileExpression);
2017-04-21 21:02:46 +02:00
while (Token::Match(tok, "} [,};]") && state.inArrayAssignment > 0U) {
tok = tok->next();
state.inArrayAssignment--;
}
} else {
state.op.push(tok);
tok = tok->tokAt(2);
}
}
}
2013-11-04 11:26:16 +01:00
}
static void compileScope(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileTerm(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "::") {
Token *binop = tok;
tok = tok->next();
if (tok && tok->str() == "~") // Jump over ~ of destructor definition
tok = tok->next();
if (tok)
compileTerm(tok, state);
if (binop->previous() && (binop->previous()->isName() || (binop->previous()->link() && binop->strAt(-1) == ">")))
compileBinOp(binop, state, nullptr);
else
compileUnaryOp(binop, state, nullptr);
2013-11-04 11:26:16 +01:00
} else break;
2012-12-16 10:06:55 +01:00
}
2013-11-04 11:26:16 +01:00
}
static bool isPrefixUnary(const Token* tok, bool cpp)
2013-11-25 20:58:40 +01:00
{
if (!tok->previous()
|| ((Token::Match(tok->previous(), "(|[|{|%op%|;|}|?|:|,|.|return|::") || (cpp && tok->strAt(-1) == "throw"))
&& (tok->previous()->tokType() != Token::eIncDecOp || tok->tokType() == Token::eIncDecOp)))
return true;
if (tok->str() == "*" && tok->previous()->tokType() == Token::eIncDecOp && isPrefixUnary(tok->previous(), cpp))
return true;
return tok->strAt(-1) == ")" && iscast(tok->linkAt(-1));
}
static void compilePrecedence2(Token *&tok, AST_state& state)
{
compileScope(tok, state);
while (tok) {
if (tok->tokType() == Token::eIncDecOp && !isPrefixUnary(tok, state.cpp)) {
compileUnaryOp(tok, state, compileScope);
} else if (tok->str() == "." && tok->strAt(1) != "*") {
if (tok->strAt(1) == ".") {
state.op.push(tok);
tok = tok->tokAt(3);
break;
} else
compileBinOp(tok, state, compileScope);
} else if (tok->str() == "[") {
if (state.cpp && isPrefixUnary(tok, state.cpp) && Token::Match(tok->link(), "] (|{")) { // Lambda
// What we do here:
// - Nest the round bracket under the square bracket.
// - Nest what follows the lambda (if anything) with the lambda opening [
// - Compile the content of the lambda function as separate tree (this is done later)
// this must be consistent with isLambdaCaptureList
2017-04-09 17:49:55 +02:00
Token* const squareBracket = tok;
if (Token::simpleMatch(squareBracket->link(), "] (")) {
Token* const roundBracket = squareBracket->link()->next();
Token* curlyBracket = roundBracket->link()->next();
while (Token::Match(curlyBracket, "%name%|.|::|&"))
2017-04-09 17:49:55 +02:00
curlyBracket = curlyBracket->next();
if (curlyBracket && curlyBracket->str() == "{") {
squareBracket->astOperand1(roundBracket);
roundBracket->astOperand1(curlyBracket);
state.op.push(squareBracket);
tok = curlyBracket->link()->next();
continue;
}
} else {
Token* const curlyBracket = squareBracket->link()->next();
squareBracket->astOperand1(curlyBracket);
state.op.push(squareBracket);
tok = curlyBracket->link()->next();
2017-04-09 17:49:55 +02:00
continue;
}
}
2017-04-09 17:49:55 +02:00
const Token* const tok2 = tok;
2017-04-09 17:49:55 +02:00
if (tok->strAt(1) != "]")
compileBinOp(tok, state, compileExpression);
else
compileUnaryOp(tok, state, compileExpression);
tok = tok2->link()->next();
} else if (tok->str() == "(" && (!iscast(tok) || Token::Match(tok->previous(), "if|while|for|switch|catch"))) {
Token* tok2 = tok;
2014-04-29 06:09:26 +02:00
tok = tok->next();
const bool opPrevTopSquare = !state.op.empty() && state.op.top() && state.op.top()->str() == "[";
const std::size_t oldOpSize = state.op.size();
compileExpression(tok, state);
tok = tok2;
2017-06-08 15:32:35 +02:00
if ((tok->previous() && tok->previous()->isName() && (!Token::Match(tok->previous(), "return|case") && (!state.cpp || !Token::Match(tok->previous(), "throw|delete"))))
|| (tok->strAt(-1) == "]" && (!state.cpp || !Token::Match(tok->linkAt(-1)->previous(), "new|delete")))
|| (tok->strAt(-1) == ">" && tok->linkAt(-1))
|| (tok->strAt(-1) == ")" && !iscast(tok->linkAt(-1))) // Don't treat brackets to clarify precedence as function calls
|| (tok->strAt(-1) == "}" && opPrevTopSquare)) {
const bool operandInside = oldOpSize < state.op.size();
if (operandInside)
compileBinOp(tok, state, nullptr);
else
compileUnaryOp(tok, state, nullptr);
}
tok = tok->link()->next();
} else if (state.cpp && tok->str() == "{" && iscpp11init(tok)) {
if (Token::simpleMatch(tok, "{ }"))
compileUnaryOp(tok, state, compileExpression);
else
compileBinOp(tok, state, compileExpression);
if (Token::simpleMatch(tok, "}"))
tok = tok->next();
2013-11-25 20:58:40 +01:00
} else break;
}
}
static void compilePrecedence3(Token *&tok, AST_state& state)
{
compilePrecedence2(tok, state);
while (tok) {
if ((Token::Match(tok, "[+-!~*&]") || tok->tokType() == Token::eIncDecOp) &&
isPrefixUnary(tok, state.cpp)) {
if (Token::Match(tok, "* [*,)]")) {
Token* tok2 = tok->next();
while (tok2->next() && tok2->str() == "*")
tok2 = tok2->next();
if (Token::Match(tok2, "[>),]")) {
tok = tok2;
continue;
}
}
compileUnaryOp(tok, state, compilePrecedence3);
} else if (tok->str() == "(" && iscast(tok)) {
Token* tok2 = tok;
tok = tok->link()->next();
if (tok && tok->str() == "(" && !iscast(tok)) {
Token *tok3 = tok->next();
compileExpression(tok3, state);
if (tok->link() == tok3)
tok = tok3->next();
} else {
compilePrecedence3(tok, state);
}
compileUnaryOp(tok2, state, nullptr);
} else if (state.cpp && Token::Match(tok, "new %name%|::|(")) {
Token* newtok = tok;
tok = tok->next();
bool innertype = false;
if (tok->str() == "(") {
if (Token::Match(tok, "( &| %name%") && Token::Match(tok->link(), ") ( %type%") && Token::simpleMatch(tok->link()->linkAt(1), ") ("))
tok = tok->link()->next();
2014-10-12 11:12:25 +02:00
if (Token::Match(tok->link(), ") ::| %type%"))
tok = tok->link()->next();
2014-10-12 12:57:01 +02:00
else if (Token::Match(tok, "( %type%") && Token::Match(tok->link(), ") [();,[]")) {
tok = tok->next();
innertype = true;
} else if (Token::Match(tok, "( &| %name%") && Token::simpleMatch(tok->link(), ") (")) {
tok = tok->next();
innertype = true;
} else {
/* bad code */
continue;
}
}
state.op.push(tok);
while (Token::Match(tok, "%name%|*|&|<|::")) {
if (tok->link())
tok = tok->link();
tok = tok->next();
}
if (Token::Match(tok, "( const| %type% ) (")) {
state.op.push(tok->next());
tok = tok->link()->next();
compileBinOp(tok, state, compilePrecedence2);
} else if (tok && (tok->str() == "[" || tok->str() == "(" || tok->str() == "{"))
compilePrecedence2(tok, state);
2014-10-12 12:57:01 +02:00
else if (innertype && Token::simpleMatch(tok, ") [")) {
tok = tok->next();
compilePrecedence2(tok, state);
}
compileUnaryOp(newtok, state, nullptr);
if (innertype && Token::simpleMatch(tok, ") ,"))
2014-10-12 10:20:03 +02:00
tok = tok->next();
} else if (state.cpp && Token::Match(tok, "delete %name%|*|&|::|(|[")) {
Token* tok2 = tok;
tok = tok->next();
if (tok && tok->str() == "[")
tok = tok->link()->next();
compilePrecedence3(tok, state);
compileUnaryOp(tok2, state, nullptr);
}
// TODO: Handle sizeof
else break;
}
}
static void compilePointerToElem(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compilePrecedence3(tok, state);
while (tok) {
if (Token::simpleMatch(tok, ". *")) {
compileBinOp(tok, state, compilePrecedence3);
} else break;
}
}
static void compileMulDiv(Token *&tok, AST_state& state)
{
compilePointerToElem(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (Token::Match(tok, "[/%]") || (tok->str() == "*" && !tok->astOperand1())) {
2014-04-26 13:32:08 +02:00
if (Token::Match(tok, "* [*,)]")) {
Token* tok2 = tok->next();
while (tok2->next() && tok2->str() == "*")
tok2 = tok2->next();
if (Token::Match(tok2, "[>),]")) {
tok = tok2;
break;
}
2014-04-26 13:32:08 +02:00
}
compileBinOp(tok, state, compilePointerToElem);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileAddSub(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileMulDiv(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (Token::Match(tok, "+|-") && !tok->astOperand1()) {
compileBinOp(tok, state, compileMulDiv);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileShift(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileAddSub(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (Token::Match(tok, "<<|>>")) {
compileBinOp(tok, state, compileAddSub);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileRelComp(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileShift(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (Token::Match(tok, "<|<=|>=|>") && !tok->link()) {
compileBinOp(tok, state, compileShift);
2013-11-04 11:26:16 +01:00
} else break;
}
}
2012-12-16 10:06:55 +01:00
static void compileEqComp(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileRelComp(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (Token::Match(tok, "==|!=")) {
compileBinOp(tok, state, compileRelComp);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileAnd(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileEqComp(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "&" && !tok->astOperand1()) {
Token* tok2 = tok->next();
2014-09-28 22:05:05 +02:00
if (!tok2)
break;
if (tok2->str() == "&")
tok2 = tok2->next();
if (state.cpp && Token::Match(tok2, ",|)")) {
tok = tok2;
break; // rValue reference
}
compileBinOp(tok, state, compileEqComp);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileXor(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileAnd(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "^") {
compileBinOp(tok, state, compileAnd);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileOr(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileXor(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "|") {
compileBinOp(tok, state, compileXor);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileLogicAnd(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileOr(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "&&") {
compileBinOp(tok, state, compileOr);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileLogicOr(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileLogicAnd(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == "||") {
compileBinOp(tok, state, compileLogicAnd);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileAssignTernary(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileLogicOr(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->isAssignmentOp()) {
state.assign++;
compileBinOp(tok, state, compileAssignTernary);
if (state.assign > 0U)
state.assign--;
} else if (tok->str() == "?") {
// 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."
2015-12-25 09:51:08 +01:00
// Hence, we rely on Tokenizer::prepareTernaryOpForAST() to add such parentheses where necessary.
if (tok->strAt(1) == ":") {
state.op.push(nullptr);
}
const unsigned int assign = state.assign;
state.assign = 0U;
compileBinOp(tok, state, compileAssignTernary);
state.assign = assign;
} else if (tok->str() == ":") {
2017-06-08 15:32:35 +02:00
if (state.depth == 1U && state.inCase)
break;
if (state.assign > 0U)
break;
compileBinOp(tok, state, compileAssignTernary);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileComma(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
compileAssignTernary(tok, state);
2013-11-04 11:26:16 +01:00
while (tok) {
if (tok->str() == ",") {
if (Token::simpleMatch(tok, ", }"))
tok = tok->next();
else
compileBinOp(tok, state, compileAssignTernary);
2013-11-04 11:26:16 +01:00
} else break;
}
}
static void compileExpression(Token *&tok, AST_state& state)
2013-11-04 11:26:16 +01:00
{
if (state.depth > AST_MAX_DEPTH)
return; // ticket #5592
2013-11-10 15:51:33 +01:00
if (tok)
compileComma(tok, state);
2013-11-04 11:26:16 +01:00
}
2016-11-20 15:44:20 +01:00
static bool isLambdaCaptureList(const Token * tok)
{
// a lambda expression '[x](y){}' is compiled as:
// [
2017-04-09 17:49:55 +02:00
// `-( <<-- optional
// `-{
// see compilePrecedence2
if (tok->str() != "[")
return false;
2017-04-10 23:29:15 +02:00
if (!Token::Match(tok->link(), "] (|{"))
return false;
if (Token::simpleMatch(tok->astOperand1(), "{") && tok->astOperand1() == tok->link()->next())
2017-04-09 17:49:55 +02:00
return true;
if (!tok->astOperand1() || tok->astOperand1()->str() != "(")
return false;
const Token * params = tok->astOperand1();
if (!params || !params->astOperand1() || params->astOperand1()->str() != "{")
return false;
return true;
}
static Token * createAstAtToken(Token *tok, bool cpp);
// Compile inner expressions inside inner ({..}) and lambda bodies
static void createAstAtTokenInner(Token * const tok1, const Token *endToken, bool cpp)
{
for (Token *tok = tok1; tok && tok != endToken; tok = tok ? tok->next() : nullptr) {
2017-04-10 23:29:15 +02:00
if (tok->str() == "{" && !iscpp11init(tok)) {
2017-04-21 21:02:46 +02:00
if (Token::simpleMatch(tok->astOperand1(), ","))
continue;
if (Token::simpleMatch(tok->previous(), "( {"))
;
// struct assignment
else if (Token::simpleMatch(tok->previous(), ") {") && Token::simpleMatch(tok->linkAt(-1), "( struct"))
continue;
// Lambda function
else if (Token::simpleMatch(tok->astParent(), "(") &&
Token::simpleMatch(tok->astParent()->astParent(), "[") &&
tok->astParent()->astParent()->astOperand1() &&
tok == tok->astParent()->astParent()->astOperand1()->astOperand1())
;
else {
// function argument is initializer list?
const Token *parent = tok->astParent();
while (Token::simpleMatch(parent, ","))
parent = parent->astParent();
if (!parent || !Token::Match(parent->previous(), "%name% ("))
// not function argument..
continue;
}
if (Token::simpleMatch(tok->previous(), "( { ."))
break;
const Token * const endToken2 = tok->link();
for (; tok && tok != endToken && tok != endToken2; tok = tok ? tok->next() : nullptr)
tok = createAstAtToken(tok, cpp);
} else if (tok->str() == "[") {
if (isLambdaCaptureList(tok)) {
2017-04-09 17:49:55 +02:00
tok = const_cast<Token *>(tok->astOperand1());
if (tok->str() == "(")
tok = const_cast<Token *>(tok->astOperand1());
const Token * const endToken2 = tok->link();
for (; tok && tok != endToken && tok != endToken2; tok = tok ? tok->next() : nullptr)
tok = createAstAtToken(tok, cpp);
}
}
}
}
2017-04-17 22:16:02 +02:00
static Token * findAstTop(Token *tok1, Token *tok2)
{
for (Token *tok = tok1; tok && (tok != tok2); tok = tok->next()) {
if (tok->astParent() || tok->astOperand1() || tok->astOperand2())
return const_cast<Token *>(tok->astTop());
if (Token::simpleMatch(tok, "( {"))
tok = tok->link();
}
for (Token *tok = tok1; tok && (tok != tok2); tok = tok->next()) {
if (tok->isName() || tok->isNumber())
return tok;
if (Token::simpleMatch(tok, "( {"))
tok = tok->link();
}
return nullptr;
}
static Token * createAstAtToken(Token *tok, bool cpp)
{
if (Token::simpleMatch(tok, "for (")) {
Token *tok2 = skipDecl(tok->tokAt(2));
Token *init1 = nullptr;
Token * const endPar = tok->next()->link();
while (tok2 && tok2 != endPar && tok2->str() != ";") {
if (tok2->str() == "<" && tok2->link()) {
tok2 = tok2->link();
if (!tok2)
break;
} else if (Token::Match(tok2, "%name% %op%|(|[|.|:|::") || Token::Match(tok2->previous(), "[(;{}] %cop%|(")) {
init1 = tok2;
AST_state state1(cpp);
compileExpression(tok2, state1);
if (Token::Match(tok2, ";|)"))
break;
init1 = nullptr;
}
if (!tok2) // #7109 invalid code
return nullptr;
tok2 = tok2->next();
}
if (!tok2 || tok2->str() != ";") {
if (tok2 == endPar && init1) {
tok->next()->astOperand2(init1);
tok->next()->astOperand1(tok);
}
return tok2;
}
Token * const init = init1 ? init1 : tok2;
Token * const semicolon1 = tok2;
tok2 = tok2->next();
AST_state state2(cpp);
compileExpression(tok2, state2);
Token * const semicolon2 = tok2;
if (!semicolon2)
return nullptr; // invalid code #7235
tok2 = tok2->next();
AST_state state3(cpp);
if (Token::simpleMatch(tok2, "( {")) {
state3.op.push(tok2->next());
tok2 = tok2->link()->next();
}
compileExpression(tok2, state3);
if (init != semicolon1)
semicolon1->astOperand1(const_cast<Token*>(init->astTop()));
tok2 = findAstTop(semicolon1->next(), semicolon2);
if (tok2)
semicolon2->astOperand1(tok2);
tok2 = findAstTop(semicolon2->next(), endPar);
if (tok2)
semicolon2->astOperand2(tok2);
else if (!state3.op.empty())
semicolon2->astOperand2(const_cast<Token*>(state3.op.top()));
semicolon1->astOperand2(semicolon2);
tok->next()->astOperand1(tok);
tok->next()->astOperand2(semicolon1);
createAstAtTokenInner(endPar->link(), endPar, cpp);
return endPar;
}
if (Token::simpleMatch(tok, "( {"))
return tok;
if (Token::Match(tok, "%type% <") && !Token::Match(tok->linkAt(1), "> [({]"))
return tok->linkAt(1);
if (Token::Match(tok, "return|case") || (cpp && tok->str() == "throw") || !tok->previous() || Token::Match(tok, "%name% %op%|(|[|.|::|<|?|;") || Token::Match(tok->previous(), "[;{}] %cop%|++|--|( !!{")) {
if (cpp && (Token::Match(tok->tokAt(-2), "[;{}] new|delete %name%") || Token::Match(tok->tokAt(-3), "[;{}] :: new|delete %name%")))
tok = tok->previous();
Token * const tok1 = tok;
AST_state state(cpp);
compileExpression(tok, state);
const Token * const endToken = tok;
if (endToken == tok1 || !endToken)
return tok1;
createAstAtTokenInner(tok1->next(), endToken, cpp);
return endToken->previous();
}
return tok;
}
void TokenList::createAst()
{
for (Token *tok = mTokensFrontBack.front; tok; tok = tok ? tok->next() : nullptr) {
tok = createAstAtToken(tok, isCPP());
}
}
void TokenList::validateAst() const
{
// Check for some known issues in AST to avoid crash/hang later on
std::set < const Token* > safeAstTokens; // list of "safe" AST tokens without endless recursion
for (const Token *tok = mTokensFrontBack.front; tok; tok = tok->next()) {
// Syntax error if binary operator only has 1 operand
if ((tok->isAssignmentOp() || tok->isComparisonOp() || Token::Match(tok,"[|^/%]")) && tok->astOperand1() && !tok->astOperand2())
throw InternalError(tok, "Syntax Error: AST broken, binary operator has only one operand.", InternalError::AST);
// Syntax error if we encounter "?" with operand2 that is not ":"
if (tok->astOperand2() && tok->str() == "?" && tok->astOperand2()->str() != ":")
throw InternalError(tok, "Syntax Error: AST broken, ternary operator lacks ':'.", InternalError::AST);
// Check for endless recursion
2018-05-09 12:36:58 +02:00
const Token* parent = tok->astParent();
if (parent) {
std::set < const Token* > astTokens; // list of anchestors
astTokens.insert(tok);
do {
if (safeAstTokens.find(parent) != safeAstTokens.end())
break;
if (astTokens.find(parent) != astTokens.end())
throw InternalError(tok, "AST broken: endless recursion from '" + tok->str() + "'", InternalError::AST);
astTokens.insert(parent);
} while ((parent = parent->astParent()) != nullptr);
safeAstTokens.insert(astTokens.begin(), astTokens.end());
2018-05-09 12:36:58 +02:00
} else if (tok->str() == ";") {
safeAstTokens.clear();
} else {
safeAstTokens.insert(tok);
2018-05-09 12:36:58 +02:00
}
// Check binary operators
if (Token::Match(tok, "%or%|%oror%|%assign%|%comp%")) {
// Skip lambda captures
if (Token::Match(tok, "= ,|]"))
continue;
// Dont check templates
if (tok->link())
continue;
// Skip pure virtual functions
if (Token::simpleMatch(tok->previous(), ") = 0"))
continue;
// Skip operator definitions
if (Token::simpleMatch(tok->previous(), "operator"))
continue;
// Skip incomplete code
if (!tok->astOperand1() && !tok->astOperand2() && !tok->astParent())
continue;
// Skip lambda assignment and/or initializer
if (Token::Match(tok, "= {|^|["))
continue;
if (!tok->astOperand1() || !tok->astOperand2())
throw InternalError(tok, "Syntax Error: AST broken, binary operator '" + tok->str() + "' doesn't have two operands.", InternalError::AST);
}
}
}
const std::string& TokenList::file(const Token *tok) const
{
2018-06-16 16:23:55 +02:00
return mFiles.at(tok->fileIndex());
}
std::string TokenList::fileLine(const Token *tok) const
{
return ErrorLogger::ErrorMessage::FileLocation(tok, this).stringify();
}
bool TokenList::validateToken(const Token* tok) const
{
if (!tok)
return true;
for (const Token *t = mTokensFrontBack.front; t; t = t->next()) {
if (tok==t)
return true;
}
return false;
}
void TokenList::simplifyStdType()
{
for (Token *tok = front(); tok; tok = tok->next()) {
if (Token::Match(tok, "char|short|int|long|unsigned|signed|double|float") || (mSettings->standards.c >= Standards::C99 && Token::Match(tok, "complex|_Complex"))) {
bool isFloat= false;
bool isSigned = false;
bool isUnsigned = false;
bool isComplex = false;
unsigned int countLong = 0;
Token* typeSpec = nullptr;
Token* tok2 = tok;
for (; tok2->next(); tok2 = tok2->next()) {
if (tok2->str() == "long") {
countLong++;
if (!isFloat)
typeSpec = tok2;
} else if (tok2->str() == "short") {
typeSpec = tok2;
} else if (tok2->str() == "unsigned")
isUnsigned = true;
else if (tok2->str() == "signed")
isSigned = true;
else if (Token::Match(tok2, "float|double")) {
isFloat = true;
typeSpec = tok2;
} else if (mSettings->standards.c >= Standards::C99 && Token::Match(tok2, "complex|_Complex"))
isComplex = !isFloat || tok2->str() == "_Complex" || Token::Match(tok2->next(), "*|&|%name%"); // Ensure that "complex" is not the variables name
else if (Token::Match(tok2, "char|int")) {
if (!typeSpec)
typeSpec = tok2;
} else
break;
}
if (!typeSpec) { // unsigned i; or similar declaration
if (!isComplex) { // Ensure that "complex" is not the variables name
tok->str("int");
tok->isSigned(isSigned);
tok->isUnsigned(isUnsigned);
}
} else {
typeSpec->isLong(typeSpec->isLong() || (isFloat && countLong == 1) || countLong > 1);
typeSpec->isComplex(typeSpec->isComplex() || (isFloat && isComplex));
typeSpec->isSigned(typeSpec->isSigned() || isSigned);
typeSpec->isUnsigned(typeSpec->isUnsigned() || isUnsigned);
// Remove specifiers
const Token* tok3 = tok->previous();
tok2 = tok2->previous();
while (tok3 != tok2) {
if (tok2 != typeSpec &&
(isComplex || !Token::Match(tok2, "complex|_Complex"))) // Ensure that "complex" is not the variables name
tok2->deleteThis();
tok2 = tok2->previous();
}
}
}
}
}