Since you say the code c5's every time m. or not m., I argue the m. is not involved in causing the c5.


That is what I believe. Fabio started another thread that fits more into the relevance of this bug. I believe Fabio said in that thread that he didn't think the bug I brought forward was related to his. In this other thread I believe it was stated that taking the 'm.' prefix out of the assignment would resolve his issue. This may be correct on the surface but I'd almost bet my life savings that this fix will cause a huge memory leak. If so then I pretty positive it is related to the bug I'm talking about.

I hope he tested that.