EIP-1193 updates

[JunichiSugiura]

Description

Currently, if we instantiate Web3 with EIP1193 compatible provider, the provider resolver resolves CustomProvider for Eth and Shh module. This causes that the provider's send method receives payload as the first argument and callback function as the second argument when a dapp calls web3.eth.getAccounts() for example.

Code to reproduce

const Web3 = require('web3')
const EventEmitter = require('events')

class Provider extends EventEmitter {
  constructor() {
    super()
    this.isEIP1193 = true
  }
}

const provider = new Provider()
const web3 = new Web3(provider)

console.log(web3.currentProvider.constructor.name)
// -> Web3EthereumProvider

console.log(web3.eth._currentProvider.constructor.name)
// -> CustomProvider

I added another condition for the resolver to check if the provider is Web3EthereumProvider then return the provider itself.

I hope this helps.

Type of change

• Bug fix
• New feature
• Breaking change
• Enhancement

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 warnings.
• I have updated or added types for all modules I've changed
• Any dependent changes have been merged and published in downstream modules.
• I ran npm run test in the root folder with success and extended the tests if necessary.
• I ran npm run build in the root folder and tested it in the browser and with node.
• I ran npm run dtslint in the root folder and tested that all my types are correct
• I have tested my code on an ethereum test network.

[coveralls]

Coverage increased (+0.003%) to 95.865% when pulling 406bf68 on LukeSugiura:hotfix/provider-resolver into 9f95015 on ethereum:1.0.

[JunichiSugiura]

Another solution is to add `isEIP1193' property to Web3EthereumProvider class. @nivida What do you think?

[JunichiSugiura]

@nivida I applied your change request and added a test for the resolver logic. Could you review this again?