chore: discv5 move #1818
chore: discv5 move #1818
Conversation
SionoiS
force-pushed
the
chore-discv5-move
branch
2 times, most recently
from
80ad5fb
to
024e08b
Compare
SionoiS
force-pushed
the
chore-discv5-move
branch
from
8e17857
to
7cb61bc
Compare
SionoiS
force-pushed
the
chore-discv5-move
branch
4 times, most recently
from
5ca3c18
to
eb6e41e
Compare
| @@ -185,6 +158,63 @@ proc findRandomPeers*(wd: WakuDiscoveryV5, pred = none(WakuDiscv5Predicate)): Fu | |||
|
|
|||
| return discoveredRecords | |||
|
|
|||
| proc searchLoop*(wd: WakuDiscoveryV5, peerManager: PeerManager, record: Option[enr.Record]) {.async.} = | |||
There was a problem hiding this comment.
Mmm. Not sure I would have this loop be part of waku_discv5. Running a discovery loop on an instantiated discv5 instance, feeding discovered records to the peer manager and bringing all of these different components together, belong to the app IMO. I know we're trying to keep app.nim clean, but I'd say that waku_discv5 should not have knowledge of the peer manager.
There was a problem hiding this comment.
Every test rely on discv5 having a search loop and before that it was node. app is not tested the same way.
The search loop testing would not be very uniform.
Also, even if app is the place to tie peer manager and discv5 the way it's done is not good IMO we should change it later.
I propose to add a comment & Issue for a later refactor of this part.
SionoiS
force-pushed
the
chore-discv5-move
branch
2 times, most recently
from
b65e0e4
to
eb6e41e
Compare
SionoiS
force-pushed
the
chore-discv5-move
branch
from
8996800
to
4fdd0e9
Compare
- Refactor discv5 start, stop & loop. - Fix tests.
SionoiS
force-pushed
the
chore-discv5-move
branch
from
4fdd0e9
to
28bf211
Compare
Description
Move discv5 out of node and into app
Changes
track #1812