GH-30863: [JS] Use a singleton StructRow proxy handler

[swallez]

Rationale for this change

Fixes #30863 by using a singleton proxy handler in StructRow's constructor. Since the handler is stateless, there is no need to create a new instance for each row.

What changes are included in this PR?

Refactoring StructRow's constructor to extract the proxy handler.

Are these changes tested?

No additional tests since this is an internal refactoring, but yarn test runs successfully.

Are there any user-facing changes?

No.

• GitHub Issue: [JS] Use a flywheel for struct row #30863

[github-actions]

⚠️ GitHub issue #30863 has been automatically assigned in GitHub to PR creator.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit b754d5a.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 37 possible false positives for unstable benchmarks that are known to sometimes produce them.

[raulcd]

@github-actions crossbow submit verify-rc-source-integration-linux-ubuntu-22.04-amd64

[github-actions]

Revision: 71590cd

Submitted crossbow builds: ursacomputing/crossbow @ actions-4738f42be4

Task	Status
verify-rc-source-integration-linux-ubuntu-22.04-amd64

[raulcd]

I am not sure if:

• [Integration][CI] verify-rc-source-integration-linux-ubuntu-22.04-amd64 is failing around JS #44294

Is related but the failures seem to be around JS integration and this is the only commit that has been introduced lately (the commit doesn't seem like has anything to do with the failures to be honest but worth pointing out)

[domoritz]

All the ci checks passed in this pull request so shouldn't this indicate that this change is not the culprit and since it's the last change from js that the issue is somewhere else?

[raulcd]

Arrow's CI is quite extended and, for example, integration tests are not run on normal CI but as nightly jobs so sometimes things are merged and they break on the nightlies. It's not ideal but a compromise as resources for GH are limited and we struggled in the past with extreme queues.
This job has started failing in the last nightlies. I do not think this PR is the culprit but the integration jobs fail on Ubuntu and in general always around JS (either producing or consuming)

[domoritz]

I see. Thanks for the additional context. I don't have cycles to debug the ci right now and we didn't make any changes to js recently besides this change.