Make contract deployments more reliable

[nambrot]

SDK Deployments are very naive today. They should be:

• Better about gas pricing (EIP1559, bump gas when spikes happen)
• More resilient to intermittent RPC issues and errors
• Resumable
• Reusable (i.e. use some contracts from a previous deploy)
• More transparent (i.e. know what is going on for the developer)

[nambrot]

Another nice thing would be to specify things like

• Retries
• gasLimit buffer
• Logging/log levels

Though I wonder if this should actually be done at the provider level instead of our abstraction level

[nambrot]

Thinking about gasLimit/estimation, it occurred to me that our provider abstraction isn't super consistent, some configuration we basically push down to the provider that is passed in in the TransactionConfig/IChainConnection, some we expose at the multiprovider/chain-connection level (supports1599) for example (which uses TransactionConfig/IChainConnection).

It makes sense why we arrived here (confirmations cannot be supported at the provider level), but should all eip1559/gas pricing/gas limit setting be done at the multiprovider level?

[nambrot]

Hmm interestingly, supposedly ethers-rs should already support automatic 1559 detection ethers-io/ethers.js#1610

[jmrossy]

Hmm interestingly, supposedly ethers-rs should already support automatic 1559 detection

@nambrot Are you confusing ethers-rs with ethers.js here?

[nambrot]

Sorry I meant ethers js

[nambrot]

Closing this for now, if further improvements should be made, file another ticket