Remove PBC based cryptography

[xandkar]

fixes #70

[vihu]

Nice work and good riddance tpke/pbc (they will not be missed!)

It would be good to run this branch with miner and do a local run.sh, start stop some nodes, ensure that things come back alive and progress is made

[xandkar]

There's still dangling code that remains. Now that curve is only ever 'BLS12-381' - it may make sense to remove it as a parameter as well? Possible future support for more curves can be structured differently anyway (as an implementation of some common interface, for instance).

Thoughts @vihu @Vagabond @evanmcc ?

[xandkar]

Nice work and good riddance tpke/pbc (they will not be missed!)

It would be good to run this branch with miner and do a local run.sh, start stop some nodes, ensure that things come back alive and progress is made

Thanks! Will do so.

[Vagabond]

Yes we can probably remove the curve parameter now.

[xandkar]

Yes we can probably remove the curve parameter now.

Removed 5ab4d22

[Vagabond]

When can the key share be undefined?

[xandkar]

@Vagabond Great question. I simply replaced:

-type key_share() :: undefined | tc_key_share:tc_key_share() | tpke_privkey:privkey().
...
key_share :: key_share(),

with

key_share :: undefined | tc_key_share:tc_key_share(),

without checking if undefined is still needed at all.

Will check.

[xandkar]

@Vagabond Indeed it doesn't look possible, since init requires it and it never appears to be removed.

Do we (still) need have_key/1 then?

[xandkar]

It appears to be used in miner as some extra sanity assertion:
https://github.com/helium/miner/blob/master/src/miner_consensus_mgr.erl#L396
https://github.com/helium/miner/blob/master/src/miner_consensus_mgr.erl#L924
https://github.com/helium/miner/blob/master/src/miner_consensus_mgr.erl#L1041

[vihu]

Looks good to me. All tests pass locally for me and have also ran run.sh in miner with this branch with some basic start/stop testing.