Another iterator bug

Browsing source code again a found this


#if defined(__x86_64__)
for(long int i = 0; (it+i) != text.end(); i++)
#else
for(int i = 0; (it+i) != text.end(); i++)
#endif
{
tempcp = (it+i).getCharacter();
tempx += mTextHelper->getGlyphWidth(tempcp);
if( tempx > (textArea.x + textArea.width) )
{
pos.x = textArea.x;
pos.y += getNewlineHeight();
// Ignore this space, because it is used for wrapping
it.moveNext();
cp = it.getCharacter();
break;
}
if( i != 0 && mTextHelper->isSpace(tempcp) )
break; // There is another space before we need a new line
}



in Text::_setCaptionHorizontal.

The problem here is, that this piece of code sits inside a for loop:


for( it = text.begin(); it != text.end(); ++it )


which means, if it is incremented in the first code block and it is equal text.end() then, the increment part of the for statement, will push the iterator past the end, which yields undefined behavior (most likely resulting in a crash).

btw. what is that

#if defined(__x86_64__)
for(long int i = 0; (it+i) != text.end(); i++)
#else

good for? If you want the largest possible unsigned integer type, you should use std::size_t instead.

don't say where you see it

Thanks for bringing this up, I will review it. I'm not fond of having integer addition mixed with iterators, I'll probably rewrite this function.

I believe that code was added in for 64 bit compatibility. I wonder if using size_t would work for everybody, like you suggest.

The use of: std::size_t works here on Linux-AMD64. So if it also works on 32-bit platforms, we should probably remove the: #if defined(__x86_64__) section and go with that.

Mike