-
Notifications
You must be signed in to change notification settings - Fork 235
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix: remove Clone
from nonce-related APIs
#1491
base: main
Are you sure you want to change the base?
fix: remove Clone
from nonce-related APIs
#1491
Conversation
@@ -54,7 +54,7 @@ impl NonceManager for SimpleNonceManager { | |||
/// | |||
/// There is also an alternative implementation [`SimpleNonceManager`] that does not store the | |||
/// transaction count locally. | |||
#[derive(Clone, Debug, Default)] | |||
#[derive(Debug, Default)] | |||
pub struct CachedNonceManager { | |||
nonces: DashMap<Address, Arc<Mutex<u64>>>, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I see, unsure how we missed this
can we instead replace this with an Arc mutex instead? not a fan of dashmap anyway
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
From a technical standpoint, yes, that would be possible. However, I would recommend against it as it's preserving the footgun for any external NonceManager
implementations downstream crates may make.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also you wouldn't need to Arc<Mutex<_>>
, just Arc<DashMap<_, _>>
would be sufficient.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this change we could do immediately, not sure I fully get the other parts of the pr yet, so I suggest to submit that separately
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The rest of the PR is just the minimal changes to fix all the errors from removing Clone
.
This PR starts by removing Clone
from CachedNonceManager
, SimpleNonceManager
, and NonceManager
.
This means none of them fulfill TxFiller
until we remove Clone
from that (and replace it with Sized
since it still needs that).
Doing that breaks the ProviderLayer
implementation for JoinFill
so we change that to move self
instead of cloning &self
.
That of course requires changing the signature of ProviderLayer::layer
itself, which then requires changing all the impl ProviderLayer
s.
The duplicate &'a
impls are just back-compat for anything which needs to call layer
by reference.
This is technically a breaking change as it modifies the public signatures of multiple traits and structs (removing clone, switching to
self
by value instead of reference), but it should have minimal impact on most downstream users.Motivation
tl;dr is cloning a
NonceManager
(particularlyCachedNonceManager
becauseDashMap
is not anArc
) is a big foot gun. It's actually even worse than that because the various layers automaticallyclone
stuff on your behalf. The why is because cloning often turns numbers that should only be used once into numbers that are used twice.Solution
Remove
Clone
from the relevant APIs, and patch up other APIs to remove their dependency on clonability ofTxFiller
s. This does result in some duplicateProviderLayer
impls (by-ref and by-val).PR Checklist
Added TestsNot applicableAdded DocumentationNot applicable