Thinking Outside the Sandbox: Violating Trust Boundaries in Uncommon Ways (paper)

From https://www.blackhat.com/us-14/arch...x-violating-trust-boundaries-in-uncommon-ways:

 

The avenues of attack may be "uncommon" but the mechanisms all involve memory allocation and pointer manipulation exploits - use-after-free, dangling pointers, stuff like that.

Kind of gives credence to the idea that safer languages should be used for userspace programs. e.g if the browsers were written in a language like Rust or Google Go - or Ada 95, for that matter - dangling pointers would be a nonissue.