[optimization] Use individual methods of async

[dignifiedquire]

Known culprits:

• this module, fixed in 255f6d1
• refactor: replace async with light weight alternatives js-ipfs-block-service#20
• Replace async with run-parallel. form-data/form-data#204
• Use modular async js-ipfs-bitswap#14

[daviddias]

When I read the title of this issue again my reaction was, "wait, no more async calls?" :D

async is now modular, not sure if this is an optimization/concern anymore

[dignifiedquire]

That is correct, we can now migrate the rest of the modules that use async to use the modular 2.0 version.

[daviddias]

Awesome :) Closing this one, as that is just a 'dep update', it will happen with following releases, naturally.

[dignifiedquire]

Sorry, but it's more than that, we have to specifically change our code to use the form of

const each = require('async/each')

to actually get the benefits of this.

[daviddias]

Ok. Can you list out all the cases where async happens (+ run-*) so that we make sure to tackle all of them? Thank you

[dignifiedquire]

I'm pretty sure that's nearly every single repo of libp2p/js-* and ipfs/js-*

[nginnever]

async crawl >.<

actually not that many modules depend on async anymore

• js-ipfs-api PR
• js-ipfs-repo (dev only) PR
• js-libp2p-kad-routing PR
• js-ipfs-unixfs-engine (dev only) PR

I have set up an example pr, if it looks good i'll iterate over the rest.

Do we care to update this if async is only a dev dep? adding them to the list for now.

[dignifiedquire]

I think we should do it for dev as well, as that means the bundles in the test runs are smaller, making it a bit faster :)

[daviddias]

Seems that this has been taken care of with exception of bitswap which has its own issue ipfs/js-ipfs-bitswap#14. Closing this one, reopen if needed