May 16, 2024

Don't think that'll work given the source and config changes. Likely we would have to pass the sodium detection to even get to the calls under the assumption that tweetnacl provides interface and initialization parity. Given that there's conditions that involve differentiated initialization calls for windows I think source changes in zmq will be necessary

Can you post the link to the zmq issue/pr where they removed it?

I see, it's not the Tweet NaCl license that's at issue. A few years back they solicited authorization for all contributors to grant relicensing (I granted it years ago). It was the guy who originally added it, I recognized him from the commit, so they pulled out that work.

Jeesh, they are still carrying and updating my "/deprecated" msvc files (all the way back to vs2010).

The commit is almost entirely build system. Seems like the actual C/C++ source change is compatible with the header swap. But yeah, lots of build system shit.