Fix burst only #396
Fix burst only #396
Conversation
This should fix some annoying log spam and rekey quicker.
Should fix bogus logging on nodes with darknet peers.
This fixes logger errors due to trying to reconnect to old opennet peers which are not in the routing table and not in the status table either, but get set to BURSTING on non-NATed nodes.
Mainly for squashing log errors, but the peer counts *are* used for important things, notably determining whether we need more opennet peers.
Conflicts: src/freenet/node/PeerNode.java
This reverts commit f98ac58. This is wrong. There is a better way... Conflicts: src/freenet/node/PeerNode.java
This makes much more sense. It also fixes the "old opennet peers in the peer status tracker" problem more cleanly, and removes some unnecessary argument cruft. (Renames forceCancelDisconnecting to onAdded and always calls it in PeerManager.addPeer)
|
I've added a bunch more fixes, I think this is reasonable now. Please re-review. :) |
|
Yes, DarknetPeerNode's setStatus method calls PeerNode's setStatus and then changes the value in some cases. I will add some javadocs. |
|
I'll try not to add more unrelated bugfixes to this branch! Please review. Thanks. |
Fixes a minor regression in recent fixes.
(Fixes more log spam)
|
Anyone want to review this? |
|
@toad could you explain what the non-logging changes here do? |
|
Looks good to me |
|
Can I trouble someone to resolve the merge conflicts? I attempted fixing them but it wasn't exceedingly clear to me what was intended. |
|
Ping on those merge conflicts. |
|
Cough cough. |
|
ping on the merge conflicts from my side, too |
|
ping? |
ArneBab
added
the
mergeAfterResolvingConflicts
|
@toad ping on the merge conflicts again |
|
Here’s a version with fixed merge errors. It needs re-review, though, because the code it patches changed quite a bit since 2016: #991 |
|
replaced by #991 |
Various cleanups, mostly directed at fixing logging errors. Relatively low risk but could conceivably change connectivity behaviour on non-NATed nodes.