fix: wake future after buffering pong

[Rjected]

This issue was uncovered in hive tests, where hive sends a single Ping after the eth handshake, and waits for a Pong response before sending any other requests over p2p. When this happens, we don't have any messages to send, so we never poll_flush the Pong out of the stream.

To fix this, we call the waker, so the Sink can be polled at least once after the Stream returns Poll::Pending.

[codecov]

Codecov Report

Merging #4536 (9bc5066) into main (0fed707) will decrease coverage by 0.01%.
Report is 6 commits behind head on main.
The diff coverage is 0.00%.

Files Changed	Coverage Δ
crates/net/eth-wire/src/p2pstream.rs	78.92% <0.00%> (-0.40%)	⬇️

... and 13 files with indirect coverage changes

Flag	Coverage Δ
integration-tests	16.69% <0.00%> (-0.03%)	⬇️
unit-tests	64.00% <0.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
reth binary	30.81% <ø> (ø)
blockchain tree	83.58% <ø> (ø)
pipeline	90.56% <ø> (ø)
storage (db)	75.32% <ø> (ø)
trie	94.84% <ø> (-0.04%)	⬇️
txpool	47.99% <ø> (-0.03%)	⬇️
networking	77.37% <0.00%> (-0.11%)	⬇️
rpc	57.43% <ø> (-0.01%)	⬇️
consensus	63.40% <ø> (ø)
revm	31.73% <ø> (ø)
payload builder	6.34% <ø> (ø)
primitives	86.53% <ø> (+0.04%)	⬆️