Fix 3380: Add AbstractProvider

[odanado]

Description

Fixes #3380

The only method that provider should implement is the send function.
I'd like you to consider whether the interface of EventEmitter should be implemented or not.

Type of change

• New feature (non-breaking change which adds functionality)

Checklist:

• I have selected the correct base branch.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• Any dependent changes have been merged and published in downstream modules.
• I ran npm run dtslint with success and extended the tests and types if necessary.
• I ran npm run test:unit with success.
• I have executed npm run test:cov and my test cases do cover all lines and branches of the added code.
• I ran npm run build-all and tested the resulting file/'s from dist folder in a browser.
• I have updated the CHANGELOG.md file in the root folder.
• I have tested my code on the live network.

[odanado]

I would like to consider the following points

• Using the extends EventEmitter.
• Remove extends EventEmitter (Leaving only the send function in the interface).
• Remove extends EventEmitter and directly define the on, once, and removeListener functions as well as the WebsocketProviderBase.

[smithki]

My $0.02 is to separate the core functionality from the subscription functionality. Not all custom providers support subscription features by default. Classes can implement multiple interfaces in TypeScript, so it shouldn't be problematic to export two interfaces.

I am not an active contributor to this code-base. This is my opinion only.

[ryanio]

Hey @odanado thanks for this PR :) It LGTM. For your 3 points, perhaps just on/once/removeListener is sufficient as the "bare minimum".

What are the types of WebsocketProviderBase?

Also, in light of recent updates to EIP-1193, there is a new request method that would be great to add. I don't think having sendAsync would hurt either even though it's deprecated it has more long term support.

[odanado]

@ryanio Thank you for the comment.
I didn't know the request method of EIP-1193.

I thought that the interface of the AbstractProvider should be an only of send and sendAsync.
The reasons are as follows.

• The status of EIP-1193 is a draft.
• EIP-1193 is not yet fully implemented in Metamask

I think it should define a provider that has methods such as on/once/removeListener/request separately from AbstractProvider.

[coveralls]

Coverage remained the same at 86.048% when pulling 20491df on odanado:fix-3380 into 67ea9d0 on ethereum:1.x.

[ryanio]

Thanks for adding sendAsync 🙌 looks like there is a small dtslint issue.

[odanado]

@ryanio I forgot that fix a test.
I updated the test.

[ryanio]

Looks great, thanks!