Deferred memproof provisioning about holochain

[artbrock]

In this state, the only two app websocket methods which will be accepted are the two new ones: all others will fail with an error message telling the user that memproofs must be provided.

I don't think this limitation is realistic. For the app UI to perform the joining process, it almost surely needs some of the "info" methods such as: AppInfo(), DnaInfo(), and AgentInfo() (even though the agent key is not finalized yet).

[matthme]

Yes agreed, AppInfo should be callable as well and return the app status ("uninitiated" or similar) as well as the currently registered public key for this app.

[artbrock]

Questions:

1. If an app / some DNAs were partially installed, but the membrane process was never completed, is there some kind of clean up process needed?
2. If the process is attempted again at some point in the future, is there a new partial install, or do we recognize an existing partial install in progress?
3. Re: Port bindings... How does the UI know what port to use? Is this the same port for the app once fully installed? Is there any clean up of ports needed?

[matthme]

Regarding 1: I don't think so, at least not for the launcher case. Such an app would show up as a normal app in launcher. From an end-user perspective there is no difference between an app that has gone through joining and one that hasn't, the difference for them lies only in what they're presented in the app UI. It is therefore up to the end-user to uninstall such an app again if they decide that they don't want to complete the joining process.