Post #1898353

@tumdum @NohatCoder also this reminds me of a technique where, for some randomly-selected subset of like 1% of malloc() calls, you allocate at the beginning, or just before the end, of a randomly-selected page boundary, and have the page before/after marked read-only. so the analogous thing here would be: some random 1% of the time, allocate with a new page that you then mark in such a way that when it's free()'d, it's really free'd.