tpu-client-next: return receiver in scheduler::run #4454
Conversation
KirillLykov
force-pushed
the
klykov/tpu-client-next-return-receiver
branch
from
97ed5c2 to
ad3fb31
Compare
KirillLykov
changed the title
| pub fn new(keypair: &Keypair) -> Self { | ||
| let (certificate, key) = new_dummy_x509_certificate(keypair); | ||
| Self { certificate, key } | ||
| pub fn new(keypair: Option<&Keypair>) -> Self { |
There was a problem hiding this comment.
As said in the PR description, not a new change. Adding back reverted PR.
| Self { certificate, key } | ||
| } else { | ||
| let (certificate, key) = new_dummy_x509_certificate(&Keypair::new()); | ||
| Self { certificate, key } |
There was a problem hiding this comment.
could condense this using unwrap_or_else:
let keypair = keypair.unwrap_or_else(|| &Keypair::new());
let (certificate, key) = new_dummy_x509_certificate(keypair);
Self { certificate, key }
There was a problem hiding this comment.
unwrap_or_else(|| &Keypair::new()) won’t work because the new Keypair instance would not live long enough for the reference to remain valid. The only thing I can do is
let keypair = if let Some(keypair) = keypair {
keypair // Use the provided keypair
} else {
&Keypair::new() // Create a new keypair if None
};
let (certificate, key) = new_dummy_x509_certificate(keypair);This way avoiding two calls of new_dummy_x509_certificate
KirillLykov
force-pushed
the
klykov/tpu-client-next-return-receiver
branch
from
ad3fb31 to
acd49ec
Compare
|
Backports to the beta branch are to be avoided unless absolutely necessary for fixing bugs, security issues, and perf regressions. Changes intended for backport should be structured such that a minimum effective diff can be committed separately from any refactoring, plumbing, cleanup, etc that are not strictly necessary to achieve the goal. Any of the latter should go only into master and ride the normal stabilization schedule. Exceptions include CI/metrics changes, CLI improvements and documentation updates on a case by case basis. |
KirillLykov
force-pushed
the
klykov/tpu-client-next-return-receiver
branch
from
acd49ec to
1db2845
Compare
|
@bw-solana I've renamed identity/validator_identity to be always stake_identity (also in tests). So now this naming is consistent. |
KirillLykov
force-pushed
the
klykov/tpu-client-next-return-receiver
branch
from
1db2845 to
2865da8
Compare
Return receiver from scheduler::run so that it can be reused if required. This is needed to use tpu-client-next in validator because user can make an rpc call to update underlying authority, and due to the way this mechanism is implemented, we have to re-utilize the same receiver. (cherry picked from commit 85b6118)
…4454) (#4521) tpu-client-next: return receiver in scheduler::run (#4454) Return receiver from scheduler::run so that it can be reused if required. This is needed to use tpu-client-next in validator because user can make an rpc call to update underlying authority, and due to the way this mechanism is implemented, we have to re-utilize the same receiver. (cherry picked from commit 85b6118) Co-authored-by: kirill lykov <[email protected]>
Problem & Solution
Return receiver from scheduler run so that it can be reused if required. This is needed to use in validator because user can make an rpc call to update underlying authority, and due to the way this mechanism is implemented, we have to reutilize the same receiver.
These changes are part of the reverted PR https://github.com/anza-xyz/agave/pull/3515/files#diff-17507818c3be9c444eb55f6c0b2d82ac0662746abf0dff6b200d43bf9ca125c2
They are split into a separate PR for the sake of backporting and to have a clearer history.