Investigate near-sandbox-rs user friction points

[r-near]

Context

The Rust DevTools status doc notes near-sandbox-rs has "high friction for new users" with documentation that's "too simple or too complex."

Action

Investigate the actual failure modes:

• Find 2-3 external users/teams who attempted to use near-sandbox-rs
• Interview or review their issues: What broke? Setup? Docs? Mental model?
• Check GitHub issues on the repo for common pain points
• Document findings

Questions to answer:

1. Are users failing to set it up at all?
2. Are they succeeding but not understanding what they did?
3. Is the friction in installation, configuration, or usage patterns?
4. How does it compare to the old near-workspaces-rs experience?

Outcome

Clear understanding of friction points to inform documentation improvements and API ergonomics.

[vsavchyn-dev]

• Singleton / shared sanbdox - apparently singleton pattern sandbox is faster than the normal model to running sandbox. It was reported to be 5 times faster, as there is no need to spin up and then remove full neard node.

  • What is done: documented new method and create example
  • What needs to be done: possibly add same support for nextest

• Set-up code redundancy / testing framework support - using near-sandbox-rs and near-api-rs togeather creates a lot of redundancy (setting up sandbox, creating network config for near-api, doing requests)

  • What needs to be done: as originally mentioned by Vlad Frolov, it would make sense to create a near-rstest or near-fixtures for reducing this duplication.

• Helper utilities - provide lightweight helpers for random account/key generation and one‑liner deployers, but retain explicit control for production‑like workflows.

  • What is done: there are already helpers for generating random account_ids and account keys, but no one-liner deployers.
  • What needs to be done: better examples that are using these random account/key generation. Possibly add easier way for generating this via near-rstest

• Lack of examples on docs - there is a lack of examples for usage of near-sandbox-rs + near-api-rs. This leaves newcomers confused how to use it and they opt-in for near-workspaces (old method)

  • What is done: right now garikbesson is doing transition of examples to near-sandbox-rs + near-api-rs
  • What needs to be done: aid garikbesson and transfer examples on near-sdk-rs from near-workspaces to near-sandbox-rs + near-api-rs

• Forking sandbox - there is a demand to fork sandbox state at a point in time that would enable test branching without full restart of the node

  • What is done: there is ability to pull and patch state from the mainnet and testnet
  • What needs to be done: possibly a NEP for "fork/revert"?