We have run much more stringent tests than those shown in your code with no issues, so I’m guessing something in the “do lots of stuff, then” block is clobbering the memory somehow so it doens’t get freed correct.y.
You’re correct, I found the problem. It was indeed an errant malloc Also, the ret issue was due to some copied C++ std::string code, so my bad on the example… I just hadn’t fixed it yet at some point. Do you have an article as to how to effectively user configASSERT() and describe what it does/how to use it in practice? I haven’t really seen a good description of this so far.