v3 roadmap

[ahdinosaur]

a boring upgrade to dogstack

• all dogstack config in index.js
• query dependency tree: v3: static query dependency tree feathers-action-react#8
• redux-cycle: update to use redux-cycles feathers-action#22
• feathers@3
  • upgrade to feathers-reactive@0.5 feathers-action#23
• feathers hook sql transactions: Service methods are not executed in transaction feathersjs-ecosystem/feathers-knex#91
• do we want to use webpack instead of browserify?
• move from uify-server to bankai
• simplify dogstack-agents: merge agents, profiles, credentials tables into agents table
• agents has property relationships: directs, manages, supplies
• upgrade to react@16
• upgrade to material-ui@next

[iainkirkpatrick]

while cleaning up bookmarks, i found this article about redux-cycle, might be useful for context

[mischa-s]

• upgrade to react@16

[danalexilewis]

get hot reloading working - (move to webpack?)

[ahdinosaur]

added #88

[danalexilewis]

Support option to use typescript. I assume this is to do with how Webpack is setup + other tsconfigs

[ahdinosaur]

@agentlewis i think it's best that v3 be a minimal set of improvements to upgrade dogstack, so no new features and mostly boring dependency upgrades. i updated the list to reflect what i think is an appropriate roadmap. feel free to make issues for wishlist features, but i strongly think we should defer as much as possible to v4, given that we aren't making much progress towards even keeping dogstack modules up-to-date.

[danalexilewis]

@ahdinosaur I get what your saying about minimal updates - that makes sense.

Can you clarify for me than that using Bankai is going to enable us to get HMR working? With something like react-hot-loader

Because otherwise it feels like we are just going to have to change the server again.

Looking at Bankai it looks like it’s mainly for the choo ecosystem and I’m kind of confused by this as dogstack is meant to be a grab bag of the popular tools - which fornme would have meant Webpack.

Can you elaborate why you have chosen to push on Bankai?

[ahdinosaur]

Can you elaborate why you have chosen to push on Bankai?

worse is better, i think bankai is the next closest step from what we have now to a better dogstack. i'd rather have something that works better than what we have now than some hypothetical ideal that nobody is actually making happen, i'd rather we decide this with working software.

i'm happy if we change to webpack in the future, by changing to bankai now we'll overcome a good chunk of the challenges that will be necessary at that point anyways, namely to split the asset and api server.