-
Notifications
You must be signed in to change notification settings - Fork 957
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
swarm/handler: replace inject_* methods #3085
Conversation
d926cd2
to
d60c0e7
Compare
on ConnectionHandler implementations
d60c0e7
to
cf9768f
Compare
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.
Cool to see this work happening. Let us know once this is ready for a review.
swarm/src/handler.rs
Outdated
#[deprecated( | ||
since = "0.41.0", | ||
note = "Handle `StreamEvent::FullyNegotiatedInbound` on `ConnectionHandler::on_event` instead. | ||
the default of implemention of this inject_*` method delegates to it." |
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 default of implemention of this inject_*` method delegates to it." | |
The default implementation of this `inject_*` method delegates to it." |
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.
Thanks Max! addressed.
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.
Cool to see this work happening. Let us know once this is ready for a review.
Thanks Max, yeah I am waiting for #3011 to be merged cause this PR also touches some of the same files to then rebase against master and mark it as ready.
swarm/src/handler.rs
Outdated
|
||
/// Enumeration with the list of the possible stream events | ||
/// to pass to [`on_event`](ConnectionHandler::on_event). | ||
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { |
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.
An alternative to explore (not saying it is a good idea) would be to take a single trait parameter, namely ConnectionHandler
here.
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { | |
pub enum StreamEvent<'a, Handler: ConnectionHandler> { |
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.
Thanks for bringing this Max,
yeah it works but we need to relax the sizedness to Handler: ConnectionHandler + ?Sized
because on_event
's argument event: StreamEvent<Self>
which afaik will probably be problematic if we in the future define methods on ConnectionHandler
that take self
and StreamEvent<Self>
cause Self
must be Sized
and StreamEvent<Self>
may be ?Sized
. Would you prefer that over the multiple generic parameters?
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.
yeah it works but we need to relax the sizedness to
Handler: ConnectionHandler + ?Sized
becauseon_event
's argumentevent: StreamEvent<Self>
which afaik will probably be problematic if we in the future define methods onConnectionHandler
that takeself
andStreamEvent<Self>
causeSelf
must beSized
andStreamEvent<Self>
may be?Sized
.
I am not sure I am following. In which case would ConnectionHandler
not be Sized
? In which case would this be problematic?
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 approach you suggest on #3085 (comment):
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { | |
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { |
implies relaxing StreamEvent<Handler>
to StreamEvent<Handler: ConnectionHandler + ?Sized>
.
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.
Looking good! Thank you :)
swarm/src/handler.rs
Outdated
/// informs the handler about an event coming from the outside in the handler. | ||
fn on_behaviour_event(&mut self, _event: Self::InEvent) {} | ||
|
||
fn on_event( |
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.
Should this be an on_connection_event
maybe? It would be nicer to have both on_
handlers use some kind of qualifier.
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.
yeah makes sense, thanks Thomas updated! :)
swarm/src/handler/map_in.rs
Outdated
match event { | ||
StreamEvent::FullyNegotiatedInbound(FullyNegotiatedInbound { protocol, info }) => | ||
{ | ||
#[allow(deprecated)] | ||
self.inner.inject_fully_negotiated_inbound(protocol, info) | ||
} | ||
StreamEvent::FullyNegotiatedOutbound(FullyNegotiatedOutbound { protocol, info }) => | ||
{ | ||
#[allow(deprecated)] | ||
self.inner.inject_fully_negotiated_outbound(protocol, info) | ||
} | ||
StreamEvent::AddressChange(AddressChange { new_address }) => | ||
{ | ||
#[allow(deprecated)] | ||
self.inner.inject_address_change(new_address) | ||
} | ||
StreamEvent::DialUpgradeError(DialUpgradeError { info, error }) => | ||
{ | ||
#[allow(deprecated)] | ||
self.inner.inject_dial_upgrade_error(info, error) | ||
} | ||
StreamEvent::ListenUpgradeError(ListenUpgradeError { info, error }) => | ||
{ | ||
#[allow(deprecated)] | ||
self.inner.inject_listen_upgrade_error(info, error) | ||
} | ||
} |
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 guess this can be replaced with self.inner.on_event
once we remove them right?
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.
yeah exactly, also see #3011 (comment)
swarm/src/handler/pending.rs
Outdated
use std::task::{Context, Poll}; | ||
use void::Void; | ||
|
||
use super::{FullyNegotiatedInbound, FullyNegotiatedOutbound, StreamEvent}; |
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'd prefer to not use super
imports.
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.
yeah, makes total sense, specailly cause we are already importing with crate::handler
. Thanks Thomas, updated!
swarm/src/handler.rs
Outdated
|
||
/// Enumeration with the list of the possible stream events | ||
/// to pass to [`on_event`](ConnectionHandler::on_event). | ||
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { |
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.
pub enum StreamEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { | |
pub enum ConnectionEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { |
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.
updated!
…n-handler-inject-on
and StreamEvent -> ConnectionEvent.
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.
Minor comments. Otherwise this looks good to me.
swarm/src/handler.rs
Outdated
#[deprecated( | ||
since = "0.41.0", | ||
note = "Handle `ConnectionEvent::FullyNegotiatedInbound` on `ConnectionHandler::on_connection_event` instead. | ||
The default of implemention of this `inject_*` method delegates to it." |
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 default of implemention of this `inject_*` method delegates to it." | |
The default implementation of this `inject_*` method delegates to it." |
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.
Thanks Max 🙇
swarm/src/handler.rs
Outdated
@@ -224,6 +277,76 @@ pub trait ConnectionHandler: Send + 'static { | |||
{ | |||
ConnectionHandlerSelect::new(self, other) | |||
} | |||
|
|||
/// informs the handler about an event coming from the outside in the handler. |
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.
/// informs the handler about an event coming from the outside in the handler. | |
/// Informs the handler about an event from the [`NetworkBehaviour`]. |
I would find this easier to understand. What do you think?
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.
yeah I agree, thanks!
swarm/src/handler.rs
Outdated
pub enum ConnectionEvent<'a, IP: InboundUpgradeSend, OP: OutboundUpgradeSend, IOI, OOI> { | ||
/// Informs the handler about the output of a successful upgrade on a new inbound substream. | ||
FullyNegotiatedInbound(FullyNegotiatedInbound<IP, IOI>), | ||
/// Informs the handler about successful upgrade on a new outbound stream. |
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.
/// Informs the handler about successful upgrade on a new outbound stream. | |
/// Informs the handler about the output of a successful upgrade on a new outbound stream. |
To be consistent with the above.
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.
yeah makes sense, thanks Max!
fn on_behaviour_event(&mut self, _event: Self::InEvent) {} | ||
|
||
fn on_connection_event( | ||
&mut self, | ||
_event: ConnectionEvent< | ||
Self::InboundProtocol, | ||
Self::OutboundProtocol, | ||
Self::InboundOpenInfo, | ||
Self::OutboundOpenInfo, | ||
>, | ||
) { | ||
} | ||
} |
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 wonder, once Self::InEvent
is called Self::FromBehaviour
, should we rename ConnectionEvent
to FromConnection
?
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.
Works for me, though I suggest not doing that within this pull request.
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.
Thanks for the follow-ups.
Happy to merge here. Let's give @thomaseizinger antoher chance to review. |
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.
LGTM!
@jxs can you take a look at the CI failure?
https://github.com/libp2p/rust-libp2p/actions/runs/3474448909/jobs/5807623337 |
This pull request has merge conflicts. Could you please resolve them @jxs? 🙏 |
thanks for noticing Max! Yeah added the missing |
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.
🚀
Description
Previously, we had one callback for each kind of message that a
ConnectionHandler
would receive from either itsNetworkBehaviour
or the connection itself.With this patch, we combine these functions, resulting in two callbacks:
on_behaviour_event
on_connection_event
Resolves #3080.
Links to any relevant issues
ConnectionHandler
counterpart to #3011Open Questions
Change checklist