walkero
/
cppcheck
1
0
Fork 0
Code Issues Pull Requests Projects Releases 1 Wiki Activity

checkFloatToIntegerOverflow: check all calculations, not only in function bodies

This commit is contained in:
Daniel Marjamäki 2017-09-19 19:30:24 +02:00
parent 9268c2034a
commit 840c7eb258
2 changed files with 51 additions and 53 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

97
lib/checktype.cpp
View File

@ -344,65 +344,60 @@ void CheckType::longCastReturnError(const Token *tok)
void CheckType::checkFloatToIntegerOverflow()
{
const SymbolDatabase *symbolDatabase = _tokenizer->getSymbolDatabase();
const std::size_t functions = symbolDatabase->functionScopes.size();
for (std::size_t i = 0; i < functions; ++i) {
const Scope * scope = symbolDatabase->functionScopes[i];
for (const Token* tok = scope->classStart->next(); tok != scope->classEnd; tok = tok->next()) {
const ValueType *vtint, *vtfloat;
const std::list<ValueFlow::Value> *floatValues;
for (const Token *tok = _tokenizer->tokens(); tok; tok = tok->next()) {
const ValueType *vtint, *vtfloat;
const std::list<ValueFlow::Value> *floatValues;
// Explicit cast
if (Token::Match(tok, "( %name%") && tok->astOperand1() && !tok->astOperand2()) {
vtint = tok->valueType();
vtfloat = tok->astOperand1()->valueType();
floatValues = &tok->astOperand1()->values();
}
// Explicit cast
if (Token::Match(tok, "( %name%") && tok->astOperand1() && !tok->astOperand2()) {
vtint = tok->valueType();
vtfloat = tok->astOperand1()->valueType();
floatValues = &tok->astOperand1()->values();
}
// Assignment
else if (tok->str() == "=" && tok->astOperand1() && tok->astOperand2()) {
vtint = tok->astOperand1()->valueType();
vtfloat = tok->astOperand2()->valueType();
floatValues = &tok->astOperand2()->values();
}
// Assignment
else if (tok->str() == "=" && tok->astOperand1() && tok->astOperand2()) {
vtint = tok->astOperand1()->valueType();
vtfloat = tok->astOperand2()->valueType();
floatValues = &tok->astOperand2()->values();
}
// TODO: function call
// TODO: function call
else
else
continue;
// Conversion of float to integer?
if (!vtint || !vtint->isIntegral())
continue;
if (!vtfloat || !vtfloat->isFloat())
continue;
for (std::list<ValueFlow::Value>::const_iterator it = floatValues->begin(); it != floatValues->end(); ++it) {
if (it->valueType != ValueFlow::Value::FLOAT)
continue;
// Conversion of float to integer?
if (!vtint || !vtint->isIntegral())
if (!_settings->isEnabled(&(*it), false))
continue;
if (!vtfloat || !vtfloat->isFloat())
continue;
for (std::list<ValueFlow::Value>::const_iterator it = floatValues->begin(); it != floatValues->end(); ++it) {
if (it->valueType != ValueFlow::Value::FLOAT)
if (it->floatValue > ~0ULL)
floatToIntegerOverflowError(tok, *it);
else if ((-it->floatValue) > (1ULL<<62))
floatToIntegerOverflowError(tok, *it);
else if (_settings->platformType != Settings::Unspecified) {
int bits = 0;
if (vtint->type == ValueType::Type::CHAR)
bits = _settings->char_bit;
else if (vtint->type == ValueType::Type::SHORT)
bits = _settings->short_bit;
else if (vtint->type == ValueType::Type::INT)
bits = _settings->int_bit;
else if (vtint->type == ValueType::Type::LONG)
bits = _settings->long_bit;
else if (vtint->type == ValueType::Type::LONGLONG)
bits = _settings->long_long_bit;
else
continue;
if (!_settings->isEnabled(&(*it), false))
continue;
if (it->floatValue > ~0ULL)
if (bits < 64 && it->floatValue >= (1ULL << bits))
floatToIntegerOverflowError(tok, *it);
else if ((-it->floatValue) > (1ULL<<62))
floatToIntegerOverflowError(tok, *it);
else if (_settings->platformType != Settings::Unspecified) {
int bits = 0;
if (vtint->type == ValueType::Type::CHAR)
bits = _settings->char_bit;
else if (vtint->type == ValueType::Type::SHORT)
bits = _settings->short_bit;
else if (vtint->type == ValueType::Type::INT)
bits = _settings->int_bit;
else if (vtint->type == ValueType::Type::LONG)
bits = _settings->long_bit;
else if (vtint->type == ValueType::Type::LONGLONG)
bits = _settings->long_long_bit;
else
continue;
if (bits < 64 && it->floatValue >= (1ULL << bits))
floatToIntegerOverflowError(tok, *it);
}
}
}
}

7
test/testtype.cpp
View File

@ -148,9 +148,9 @@ private:
}
void signConversion() {
check("x = -4 * (unsigned)y;");
check("x = -4 * (unsigned)y;");
ASSERT_EQUALS("[test.cpp:1]: (warning) Suspicious code: sign conversion of -4 in calculation because '-4' has a negative value\n", errout.str());
check("unsigned int f1(signed int x, unsigned int y) {" // x is signed
" return x * y;\n"
"}\n"
@ -235,6 +235,9 @@ private:
}
void checkFloatToIntegerOverflow() {
check("x = (int)1E100;");
ASSERT_EQUALS("[test.cpp:1]: (error) Undefined behaviour: float () to integer conversion overflow.\n", removeFloat(errout.str()));
check("void f(void) {\n"
" return (int)1E100;\n"
"}\n");