Evaluate update to Nim 2.x

[gmega]

With gc:refc at first.

[2-towns]

I will create subtasks to ensure clear tracking of comments, as I believe we will have many comments.

[2-towns]

I updated constantine dependency to the last commit hash because there is a compilation error with Nim 2.X.

Additionally, I updated the PR created by Ben for GCC 14 on nim-poseidon2 to reflect this change.

[2-towns]

There is a problem with the default SAT resolver of nimble; it fails when trying to install dependencies; see vacp2p/nim-libp2p#1219.

I will use nimble --solver:legacy install until it is fixed.

[2-towns]

With gc:refc at first.

It became mm:refc (https://nim-lang.org/docs/mm.html).

[2-towns]

Contributions to sub-dependencies:

• nim-poseidon2: bumps constantine, updates curves nim-poseidon2#21 (resume Ben's work)
• nim-datatore: Update to nim 2 x nim-datastore#75
• nim-codex-dht: Update to nim 2 x nim-codex-dht#104
• nim-eth-versioned: Update to nim 2 x nim-eth-versioned#1
• nim-ethers: Update to nim 2 x nim-ethers#103
• nim-contract-abi: Update for Nim 2.0.x nim-contract-abi#9
• nim-nitro: Update to nim 2 x nim-nitro#2

Found issue in nim-results:

arnetheduck/nim-results#50

Contribute to nim-testutils to fix case:

status-im/nim-testutils#54

[2-towns]

I disabled for now the thread analysis using --threadAnalysis:off

[2-towns]

When trying to update nim-nat-transversal, it wasn't working on Windows in CI.

I explored the issue in depth and found that it occurred due to upgrading miniupnp from version 2.2.4 to 2.2.5. The reason for this can be traced to this commit. It created a symbolic link inside the repository, and Windows handles symbolic links with Git very poorly.

The solution seems simple: apply git config --global core.symlinks false before fetching the submodules for Windows only.

[2-towns]

I updated to Nim 2.0.14.