(WIP) Feat/thread access controllers #435
Conversation
|
Still have to work out naming, mostly depends on how we address/name access controllers, and define what makes a unique access controller, thus a unique thread. Moderator access controller name should probably depend on who first mod is, and then would bubble up as different thread addresses |
|
Unique thread access controller is defined by rootMod (first mod), if members, and thread name (which is thread name, plus space). This will likely also change how threads are store in space, will want more of this data, so that you have a full reference to thread, or could store the full thread address. Can also change thread name to include this. |
| 'use strict' | ||
|
|
||
| const pMapSeries = require('p-map-series') | ||
| const AccessController = require('orbit-db-access-controllers/src/access-controller-interface') |
| // TODO should memeber and non member threads create different acess controllers | ||
| // or how to deal with threads with same name with member and non member (different names? | ||
| // address of this access controler probably should just change, then different thread name | ||
| this._db = await this._orbitdb.feed(ensureAddress(address), { |
There was a problem hiding this comment.
What happens here if address is not passed?
| } | ||
|
|
||
| /* Factory */ | ||
| static async create (orbitdb, options = {}) { |
There was a problem hiding this comment.
Seems like there are a bunch of things missing here, but I assume you have the entire model figured out?
zachferland
force-pushed
the
feat/thread-access-controllers
branch
from
4cf78fb to
7636c24
Compare
|
This first go is probably not sufficient, for our requirements seems like there needs be some causal dependence between the ACL and the data structure. One option we already discussed is to make them part of same struct, otherwise you may be able to have other pointers as well. Many tradeoffs, if separate may to have rebuild entire thread again when receiving AC update (testing each entry against updated AC), if part of the same struct, cant fetch last n entries, need entire struct to build AC. Also some links for similar problems/work gathered by ipfs for their crdt work https://github.com/ipfs/dynamic-data-and-capabilities |
|
implementation note, access controller right now is feedstore where each entry is {capability, id } But options exist here, for more fields, or could be kv store where capability is a key and value is a list of ids, but feedstore seems more efficient. |
|
Now that a thread is also uniquely defined by rootMod and if membersOnly, we could also add that to a a name of thread. Not necessary since that info is also embedded in entire db address by being included in the manifest file. But also means we should probably offer the option to open a thread by name or address and should store thread subscription as entire address or with name along with rootMod and if members. Thats all the info needed to reference it. |
…/helpers for clarity
…r opts to mod controller
|
Also need to still inherit from access controller interface to get the event emitters for updates, but some inheritance issues since one goes to es5 and the other is es6 still, need to reconcile those two things |
|
Is it worth updating the entire 3box codebase to es6 maybe? |
|
@oed for the comments here that relate to changes for the upgrade, not related to ACs, ill look at those more closely when rebasing after |
|
Most of it is, or do you mean not traspile to es5? I think the problem is that orbit-db-access-controllers is not transpiled, so our transpiled es5 code is trying to inherit from the still es6 class |
|
Oh, I see. I don't think we need to inherit from them though. I don't in the AC I wrote. |
|
I don't yet either, and is most everything, the only thing was they also inherit that class from the event emitter, so it has this.emit for emitting update events, for the when the AC gets a new entry, which I suppose revalidates the entires after |
|
Ok, I think I missed that. |
zachferland
force-pushed
the
feat/thread-access-controllers
branch
from
cd4d87f to
c89c922
Compare
zachferland
force-pushed
the
feat/thread-access-controllers
branch
from
c89c922 to
e09d8a4
Compare
|
Closed by #455 |
No description provided.