Message 267620 - Python tracker

My current patch avoids the memory peak *and* doesn't add any memory fragmentation on top of whatever is already there.

In other words, it makes the parser better in this one aspect, and it doesn't make it worse in any aspect.