cuprated: txpool #312
cuprated: txpool #312
Conversation
github-actions
bot
added
A-p2p
| return; | ||
| }; | ||
|
|
||
| // TODO: There is a race condition possible if a tx and block come in at the same time: <https://github.com/Cuprate/cuprate/issues/314>. |
There was a problem hiding this comment.
The REORG_LOCK is not enough to prevent certain situations, we will need a lock for every blockchain mutation. This shouldn't be too bad IMO, every ~2 mins needing to take a write lock, read locks will only need to be taken when multiple services are involved in handling a request,
There was a problem hiding this comment.
to be clear all this would allow is a double spend tx in the pool, and this would only happen in a very unlucky situation.
So not really an issue however IMO we should aim to be correct under all possible scenarios, being lax here might have consequence elsewhere.
|
Btw the txpool issues should probably be |
|
I'm planning to mark the issues as help wanted when this is merged - I'll remove the |
|
|
||
| /// An identifier for a P2P peer on any network. | ||
| #[derive(Debug, Clone, PartialEq, Eq, Hash)] | ||
| pub enum CrossNetworkInternalPeerId { |
There was a problem hiding this comment.
Is there a reason this is defined here instead of in cuprate-p2p?
There was a problem hiding this comment.
defining it here means we are more free to change it without it being a breaking change + this isn't used in cuprate-p2p.
| /// TODO: should we expose this? probably not. | ||
| const DANDELION_CONFIG: DandelionConfig = DandelionConfig { |
There was a problem hiding this comment.
If this is eventually exposed, it could/should be an associated constant on the type:
impl DandelionConfig {
const CUPRATED_CONFIG: Self = /*...*/
}There was a problem hiding this comment.
I meant exposed to users of cuprated, I'll update the comment.
What
The transaction pool