polymorphic with improved exports

[Keyrxng]

Relates to ubiquity/pay.ubq.fi#205

Original context: ubiquity/pay.ubq.fi#205 (comment)

• reliance on package.json to serve correct module from @ubiquitydao/rpc-handler
• added a fallback hook that handles dynamic module export if ^ has issues
• added rpcTimeout as config option
• added test-specific build (replacing axios with fetch due to axios adapter issues)
• added axios back as it's err handling is better than fetch (I swapped it out originally 1. to reduce pkg size and 2. to pass tests)
• fixed bug with latency naming causing chain-agnostic rpc calls, meaning regardless of network all localStorage latencies would be considered safe (at least was the case for chains 100 & 1) and executed on.
• • local storage is wiped if coming from chain A to chain B

Allowing the data-tracking to be configurable would be a substantial change and would beef the package up quite handsomely because it would require that all RPCs for every chain get injected at build time and then adding a filtering method for returning list of acceptable providers meeting their data-tracking needs

What's the consensus in adding this feature?

[github-actions]

Lines	Statements	Branches	Functions
	Unknown% (0/0)	Unknown% (0/0)	Unknown% (0/0)

JUnit

Tests	Skipped	Failures	Errors	Time
0	0 💤	0 ❌	0 🔥	3.157s ⏱️

Coverage Report (0%)

File	% Stmts	% Branch	% Funcs	% Lines	Uncovered Line #s
All files	0	0	0	0

[0x4007]

Data tracking feature is fine but I'm personally not concerned with it.

Also consider making more precise changes for handling issues. And then opening a separate pull for refactoring and miscellaneous changes.

It feels like searching for a needle in a haystack for the reviewers otherwise and it's difficult to focus on the important changes with a pull of this size.

[Keyrxng]

Data tracking feature is fine but I'm personally not concerned with it.

Also consider making more precise changes for handling issues. And then opening a separate pull for refactoring and miscellaneous changes.

It feels like searching for a needle in a haystack for the reviewers otherwise and it's difficult to focus on the important changes with a pull of this size.

duly noted! After seeing molecula handle the commit history so clean I thought best to try follow suit but I'm not quite at that level yet eh lmao

I'll refactor optionals to default tmrw but I do believe that the rpcTimeout is a feature that should stay

[Keyrxng]

Is there anything holding this back except the merge conflict? If there is I'll handle them together

[Keyrxng]

I have split this PR into two, one for tests and one for polymorphic exports.

If it's easier to review the two split then this could be closed and I'll open the other two, unless that would make further review more difficult.

Either way, considering the NPM package is now published and progress can be made in relevant issues it's better to close this as not planned and stick with import {} from "@ubq../dist/esm/src/... or go through with this as-is or through a new PR specific to improving only the exports and using import {} from @ubq/rpc-handler

[rndquu]

@Keyrxng This PR should be closed in favour of #14 and #15, correct?

[Keyrxng]

That's right @rndquu