Overzealous UaF/Double Free detection.

[impost0r]

Hi. First off want to say that this is a pretty great tool and helps cut down on research time. Like everything it's prone to false positvies or negatives but it's been doing well so far.

Except when the C++ standard library is concerned. You see, these all point to the same call.

This very egregious result all points to one function. if you just glance over HLIL, you can see what might look like a double-free or use-after-free issue.

Except for the fact that if you actually look at the HLIL, it appears to be walking the arguments and freeing them. No occurence of freeing the same argument twice, nothing that looks like (to me) a use-after-free. Then again, I'm still learning the ropes of HLIL.

I can send the binary in question, though as it's a private affair it'd have to be over email or something. Either way, I'd suggest tuning the UaF/Double free/Free "issues" in general to be a lot less zealous, and if possible, ignore C++ standard library functions, because I doubt that if this were present, it'd have made it past the C++ standards committee.

Keep up the good work, regardless. Aside from the problems with free/operator delete, it has found actual bugs and does cut down on research time. If you can provide an email to where I can share the binary (and a few others that have had similar issues) I'd be happy to help you investigate the issue.

Best regards,
impost0r

[Martyx00]

Hello,
Thanks for pointing this out. Please send the binaries you can to [email protected]. I will not have much of a free time in upcoming 3 weeks but I will try to have a look whenever I can.

[impost0r]

Sure. My apologies for the late response, I've been busy as well.