cfcohen

So... I've actually looked at this a little now (instead of just reading the commentary). I think it's pretty clear that 0x8b02e8 is in fact a virtual function table. There...

Hmm. We've already got this rule. :-( https://github.com/cmu-sei/pharos/blob/master/share/prolog/oorules/rules.pl#L2466 So the question really is: Why didn't it trigger properly, and prevent these classes from being merged? For example, Instruction 0x422312 installs...

More curious results... @sei-eschwartz ran this file to completion on Jul 3 & 4, and that run did NOT merge these methods, instead assigning them to separate classes as OOAnalyzer...

Yes. Arrays of objects were known to be a topic that requires improvement. Are you able to determine exactly what we don't handle? I assume we found almost nothing in...

I assume that this is also the same root issue: ```python import numbers from typing import Union FixedRational = Union[int, numbers.Rational] def broken(n: numbers.Rational) -> int: return n.numerator * n.denominator...

Sounds great. To be honest I didn't fully realize that the static typing annotations was a Python 3 only thing because I've been exclusively using Python3 for so long now....

Huh. I hadn't actually noticed. I'm really glad to see that you're finding value in the changes. They were fairly arbitrary, and non-trivial in scope. It seems to me that...

Since I probably know my changes better, here are the commits I think of as being primarily related to the numpy elimination: https://github.com/jjgomera/iapws/pull/53/commits/38d0a95dd12e3fb6924a79b0114dc7f7238749d3 Convert to math.exp() and math.log() where possible....