Details
-
Wish
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
None
Description
I just tried running the tests on a default build on Ubuntu 20.04 with an AMD Ryzen CPU (where presumably RDRAND can't be relied on, but I don't know if that's the underlying reason), and the tests appear to hand blocking on std::random_device at some point. I suppose some versions of libstdc++ will use /dev/random instead of /dev/urandom (which is a bad idea in itself).
We should probably try to minimize our usage of std::random_device. After all, we're not generating cryptographic keys or anything.
Attachments
Issue Links
- links to