5 hours later…
15:48
I'm trying to find a post where one of the Bitcoin developers described finding a bug in OpenSSL, which affected squaring a specific number on a specific architecture, where the developer described using a testing technique that they developed when they were working on an audio driver. Does anyone have a better memory or better google-fu?
4 hours later…
19:58
Interestingly, Greg mentions that it's through doing something similar to what they did in the Opus codec, but to the best of my memory I was actually not aware of that until he posted this. Doing the biased-random-number testing approach was something I learned from the libgmp codebase, and replicated.
« first day (1774 days earlier) ← previous day next day → last day (2526 days later) »