monorepo: upgrade ethereum-crypography and reduce reliance on it

[gabrocheleau]

This PR:

• Upgrades ethereum-cryptography the latest release (3.2.0)
• Refactors direct @noble/hashes and @noble/curves re-exports from ethereum-cryptography to directly use those rather than going through ethereum-cryptography.

I was evaluated the possibility of fully removing reliance on ethereum-cryptography, but this would require re-writing some helpers or wrapper functions (e.g. randomBytes, utf8ToBytes, byteToUtf8, equalsBytes and some others). We can consider doing that eventually if we feel strongly about it.

This PR still allows us to remove the ethereum-cryptography direct dependencies from several packages.

[codecov]

Codecov Report

Attention: Patch coverage is 95.86777% with 5 lines in your changes missing coverage. Please review.

Project coverage is 79.56%. Comparing base (f10cb80) to head (e490bb9).
Report is 1 commits behind head on master.

Additional details and impacted files

Flag	Coverage Δ
block	84.33% <100.00%> (ø)
blockchain	89.32% <ø> (ø)
client	67.99% <100.00%> (?)
common	97.51% <ø> (ø)
devp2p	86.78% <100.00%> (ø)
evm	73.11% <75.00%> (ø)
mpt	90.05% <100.00%> (+0.30%)	⬆️
statemanager	69.06% <93.33%> (ø)
static	99.11% <90.90%> (ø)
tx	90.59% <100.00%> (ø)
util	89.36% <100.00%> (ø)
vm	55.50% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[holgerd77]

Urgent:

This change substantially increases bundle sizes by double-including Noble now several times, e.g. Block going up from 324KB to 442KB:

Before:

After:

(this is actually the reason why we just should not do these kind of last minute changes, because it increases the risk of something happening)

So, this needs urgently to be investigated.

In doubt (my preference) we should revert the whole PR, and put this aside.

[paulmillr]

I was evaluated the possibility of fully removing reliance on ethereum-cryptography, but this would require re-writing some helpers or wrapper functions (e.g. randomBytes, utf8ToBytes, byteToUtf8, equalsBytes and some others). We can consider doing that eventually if we feel strongly about it.

@gabrocheleau what's up with that methods?

randomBytes, utf8ToBytes, byteToUtf8, equalsBytes

All of them are present in either noble-hashes or noble-curves. hexToBytes in E-C is a wrapper around hashes hexToBytes which adds 0x prefix. curves have constant-time equalsBytes which may be slightly less performant but it's still ok.

https://github.com/ethereum/js-ethereum-cryptography/blob/31f980b2847545d33268f2510ba38a3836202a44/src/utils.ts#L10

[gabrocheleau]

I was evaluated the possibility of fully removing reliance on ethereum-cryptography, but this would require re-writing some helpers or wrapper functions (e.g. randomBytes, utf8ToBytes, byteToUtf8, equalsBytes and some others). We can consider doing that eventually if we feel strongly about it.

@gabrocheleau what's up with that methods?

randomBytes, utf8ToBytes, byteToUtf8, equalsBytes

All of them are present in either noble-hashes or noble-curves. hexToBytes in E-C is a wrapper around hashes hexToBytes which adds 0x prefix. curves have constant-time equalsBytes which may be slightly less performant but it's still ok.

ethereum/js-ethereum-cryptography@31f980b/src/utils.ts#L10

Indeed, the wrapper-related logic is minimal, but we would still need to rewrite these wrappers (if we need them) in our library, which we felt was a bit too "last-minute" for the imminent release. Not saying we won't do it, just out of scope for this release. Also have to evaluate impact on bundle-size