Howard Wu

^These commits bring over desirable changes from the `testnet2` branch.

I think this is worth including. Question: is there a way to harmonize this feature with the design, so that we adhere to an implicit design for using upnp (as...

Closing as this feature has been removed from the `testnet3` branch. Feel free to reopen if you encounter this issue elsewhere!

I think we should remove usage of the file system from the `powers.rs` implementation. The concern is that Rust doesn't check these at compile time (except in limited fashion if...

The error stems from `.unwrap()`s which need to be removed. In addition, the code is relying on the OS to create a file-system temporary folder, and immediately begins using it....

So I've been able to work out what's happening. We are never even using the `powers_of_{degree}` files. This infra solely relies on the temp file existing in your computer to...

I've refactored `PowersOfG` to address some of the above concerns. Note the version now does not use the `powers_of_*` files at all, instead it is all from the setup. Changes...

Adding my thoughts on this thread, I think the `.bits()` and `.bytes()` proposal is better than the `Bytes::from` proposal. However, from many first-hand encounters, I think the (slightly less attractive)...

I'll add an additional proposal: We could stick with the `.bits()` and `.bytes()` convention, strictly (and explicitly) enforcing only 1 form of endianness (e.g. all calls will always return little-endian),...

That sounds neat. Do you mean you construct circuits in libsnark (using something like gadgetlib) and would find it useful to export it to DIZK for running the setup and...