feat: add pipeline mode for streaming replication

[drmingdrmer]

Changelog

docs: update getting-started guide to use RaftNetworkV2 as primary trait

Modernize the getting-started documentation to focus on RaftNetworkV2
as the recommended network trait. Add documentation for the optional
stream_append() method for pipelined replication.

refactor: use stream_append for linearizable read confirmation

Replace append_entries with stream_append in linearizable read
confirmation so applications only need to implement stream_append.

refactor: use stream_append for heartbeat instead of append_entries

Replace append_entries with stream_append in HeartbeatWorker so
applications only need to implement stream_append for replication.

docs: gRPC example: implement gRPC bidirectional streaming for pipeline replication

Replace chunked append_entries with native gRPC bidirectional streaming
via stream_append. This provides more efficient pipelined log replication.

Changes:

• Add StreamAppend RPC with bidirectional streaming to proto
• Implement stream_append server handler in raft_service.rs
• Implement stream_append client in network/mod.rs
• Remove chunked append_entries fallback logic
• Change RaftNetworkV2::stream_append to accept 'static stream
• Update README to reflect streaming pipeline approach
• Delete obsolete test_chunk.rs

feat: add pipeline mode for streaming replication

Replication to a follower has two phases:

1. Binary search phase: The leader runs a binary search to find the exact
   matching position of log entries on the follower.

2. Pipeline mode: After finding the match point, the leader calls the
   stream_append method on the network and continuously generates
   AppendEntries requests. The network implementation should pipeline all
   requests to the follower and yield responses. Note that responses and
   requests don't have to be 1-to-1 mapped - the number of responses can be
   smaller than the number of requests.

stream_append provides a default implementation that calls the existing
append_entries method to emulate streaming replication. A mature
implementation should run in real pipeline mode instead of request-response
manner. When a request is received by stream_append, it is responsible for
sending all content of the request to the follower - partial success is not
allowed.

On the leader, the Inflight structure in Engine tracks the inflight
replication session running by ReplicationCore. An InflightId identifies
each inflight session, and the Inflight structure ignores any response that
doesn't match the current InflightId.

This change also reorganizes replication data structures for clarity:

• Add Payload enum for log replication specifications (LogIdRange, LogsSince)
• Add Replicate struct combining inflight_id with Payload
• Add ReplicationProgress to track local committed and remote matched state
• Simplify drain_events() to set fields directly instead of returning values
• Remove obsolete request.rs, replication_state.rs, log_state.rs

Changes:

• Add pipeline mode entry in ProgressEntry::next_send() when fully caught up
• Add Inflight::LogsSince variant for unbounded log streaming
• Add is_logs_since() method to Inflight for type checking
• Add get_partial_success() method to AppendEntriesResponse
• Refactor drain_events() to set next_action and inflight_id directly
• Add unit tests for pipeline mode in progress/entry/tests.rs

feat: add streaming replication with I/O progress synchronization

Enable replication tasks to synchronize with leader I/O progress using
watch channels. The replication stream monitors io_accepted_rx and
io_submitted_rx to detect leader changes and wait for log availability.

feat: add watch channel for I/O acceptance notification

Add io_accepted_tx watch channel to notify observers before I/O operations
are submitted to storage. This enables preparation for upcoming I/O events
before they actually happen.

Changes:

• Add io_accepted_tx watch channel to RaftCore
• Broadcast I/O acceptance before UpdateIOProgress, AppendEntries, and SaveVote

test: add assertion for watch channel changed() behavior

Add test assertion to verify that after changed() returns ready, a
subsequent call returns pending because the value was already seen.

Changes:

• Add assertion for changed() returning pending after value is marked as seen

feat: add watch channel for I/O submission progress broadcast

Add io_submitted_tx watch channel to notify replication tasks when log
entries have been submitted to storage and are safe to read. This enables
replication tasks to coordinate with I/O progress without polling.

Changes:

• Add io_submitted_tx watch channel to RaftCore
• Broadcast I/O submission progress after AppendEntries, SaveVote, and UpdateIOProgress

chore: add doc

feat: add streaming replication support with LogsSince variant

Add infrastructure for open-ended streaming replication where the leader
continuously sends logs after a given point without a fixed upper bound.
This complements the existing fixed-range Logs replication mode.

chore: remove unused inflight-id

---

• Improvement
• Build/Testing/CI

---

This change is