Upgrade dependencies

[kigawas]

No description provided.

[kigawas]

Don't know why one check succeeded and another failed

[kigawas]

Failed one: https://travis-ci.org/KZen-networks/curv/jobs/572583063

[kigawas]

Looks like it randomly failed at check_bit_length....

[omershlo]

yes, there's an issue I think about it. it s a probabilistic test and therefore it fails once in a while

[kigawas]

@omershlo
Looks like it always worked on my dev machine and github's builtin travis

But repeatedly failed on travis-ci.org

[omershlo]

coincidence I would say.

[kigawas]

Cool, I guess it's better to remove travis-ci.org check

[kigawas]

@omershlo
Did some minor updates on this, do you have some time to review?

[omershlo]

Yes, I was planning on doing it over the weekend. My main issue is with updates to secp256k1 library. I need to make sure what changes are included in that version and why not bump it to the latest

[kigawas]

yes, there's an issue I think about it. it s a probabilistic test and therefore it fails once in a while

I guess I know why.
It should be related with prefix zeros in hex format of BigInt numbers

let l63 = BigInt::from_str_radix(
    "96208c61220f2cce171f0c1d10db7c094245fb66b17b9de3e0c4fbf4e68e73d",
    16,
)
.unwrap();
assert_eq!(l63.to_str_radix(16).len(), 63);

let l63 = BigInt::from_str_radix(
    "096208c61220f2cce171f0c1d10db7c094245fb66b17b9de3e0c4fbf4e68e73d",
    16,
)
.unwrap();
assert_eq!(l63.to_str_radix(16).len(), 63);

[kigawas]

And since HashCommitment is the BigInt format of sha256, it probably starts with zero, this check is not appropriate.

[omershlo]

This check is indeed problematic because of the probabilistic nature. Do you suggest we just remove it ?

[kigawas]

I'll remove it in #61