Right now sadness-generator just allocates a 999MiB buffer on the stack to generate a stack overflow crash, however it was pointed out that catching stack overflows in Firefox is pretty unreliable, and is tested via recursion, so using a simple buffer might not be testing stack overflow handling as effectively as I had assumed, so might be useful to add or change to a recursion approach to see if that exposes reliability issues that I haven't encountered thusfar.
Right now
sadness-generatorjust allocates a 999MiB buffer on the stack to generate a stack overflow crash, however it was pointed out that catching stack overflows in Firefox is pretty unreliable, and is tested via recursion, so using a simple buffer might not be testing stack overflow handling as effectively as I had assumed, so might be useful to add or change to a recursion approach to see if that exposes reliability issues that I haven't encountered thusfar.