[LLVMbugs] [Bug 15893] New: False positive null reference warning in C++ code
bugzilla-daemon at llvm.org
bugzilla-daemon at llvm.org
Wed May 1 18:58:58 PDT 2013
http://llvm.org/bugs/show_bug.cgi?id=15893
Bug ID: 15893
Summary: False positive null reference warning in C++ code
Product: clang
Version: trunk
Hardware: Macintosh
OS: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: Static Analyzer
Assignee: kremenek at apple.com
Reporter: bugzilla at jwwalker.com
CC: llvmbugs at cs.uiuc.edu
Classification: Unclassified
Overview:
The static analyzer says that a null reference is being returned. As far as I
can tell, the code cannot do that, though it could throw an exception.
Steps to reproduce:
Analyze this code:
-------------------------
typedef int MySizeType;
class MyArray
{
public:
MyArray();
~MyArray();
int& operator[]( int index ) { return mArray[ index ]; }
MySizeType size() const { return mSize; }
MySizeType capacity() const { return mCapacity; }
void grow( MySizeType newSize );
private:
int* mArray;
MySizeType mSize;
MySizeType mCapacity;
};
MyArray::MyArray()
: mArray( NULL )
, mSize( 0 )
, mCapacity( 0 )
{
}
MyArray::~MyArray()
{
delete [] mArray;
}
void MyArray::grow( MySizeType newSize )
{
mSize = newSize;
if (mSize > mCapacity)
{
mCapacity = mSize;
delete [] mArray;
mArray = new int[ mCapacity ];
}
}
static void foo( MyArray& outFaces )
{
outFaces.grow( 4 );
for ( unsigned int i = 0; i < 4; ++i )
{
outFaces[ 0 ] = 1;
}
}
static void bar()
{
MyArray xFaces;
MyArray facesWithCopies;
foo( facesWithCopies );
const unsigned int kNumFaces = facesWithCopies.size();
xFaces.grow( kNumFaces );
if (kNumFaces > 0)
{
xFaces[0] = 0;
}
}
-------------------------
Actual results:
warning: Returning null reference
pointing to the line of code xFaces[0] = 0.
Expected results:
No warning.
Build date:
clang version 3.3 (trunk 180897)
Target: x86_64-apple-darwin11.4.2
Thread model: posix
Additional information:
If I change the first line to
typedef unsigned int MySizeType;
then the warning goes away. I can see that it's probably a good idea to
represent a size with an unsigned type, but I don't understand what Clang is
complaining about in this particular case.
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130502/e09f0d47/attachment.html>
More information about the llvm-bugs
mailing list