fix: "Cannot mix BigInt and other types" for signing transaction with explicitly provided nonce

[frol]

Pre-flight checklist

• I have read the Contributing Guidelines on pull requests.
• Commit messages follow the conventional commits spec
• If this is a code change: I have written unit tests.
• If this changes code in a published package: I have run pnpm changeset to create a changeset JSON document appropriate for this change.
• If this is a new API or substantial change: the PR has an accompanying issue (closes #0000) and the maintainers have approved on my working plan.

Motivation

Test Plan

Related issues/PRs

[changeset-bot]

🦋 Changeset detected

Latest commit: ed4404f

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@near-js/client	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[frol]

@gagdiez May I ask for your review? We have discovered this issue today with @BenKurrek since he was manually controlling nonce and NAJ was crashing his app since undefined += 1n did not work.

[gagdiez]

why not just initializing the nonce as 0n? I'm quite sure that leaving it as undefined will break something further down the line, specially with borsh serialization

What is the issue / code that originated this PR?

[frol]

@gagdiez I considered that, but initializing it with 0 will create other bugs - if you first sign a transaction with explicit nonce and next time without explicit nonce, it will use nonce=1, which is wrong. Also, just on the conceptual level it is not a right move to initialize it with invalid value.

Regarding you concern that this fix might bite us elsewhere, I don’t believe so as this value is tied to this internal state and not used directly anywhere else.

[gagdiez]

I have no better solution for now, so I will not block this, approved