fix: add timeouts to handshake I/O operations (#11)

[rafabd1]

Fix: Add timeouts to handshake I/O operations

Closes #11

Problem

When I2P tunnels are degraded (frequent during bootstrapping or network instability), peer connection establishment hangs indefinitely because the handshake I/O operations (read_framed / write_framed) have no timeouts.

The I2P streaming layer accepts the inbound connection, but the application-level handshake never completes because data can't flow through broken tunnels — and nothing ever times out to report the failure.

Solution

Added timeouts to all handshake I/O operations:

1. In handle_incoming (responder side):

   • Read HANDSHAKE_INIT with 60-second timeout (given I2P latency)
   • Send HANDSHAKE_ACK with 30-second timeout (writing should be fast)

2. In initiate_session (initiator side):

   • Send HANDSHAKE_INIT with 30-second timeout
   • Wait for HANDSHAKE_ACK with 60-second timeout

3. Emit user-visible errors:

   • When handshake fails (timeout or otherwise), emit a connection_error event with a descriptive message
   • Users now see "Connection failed: handshake timeout: peer did not send INIT within 60s" instead of indefinite hanging

Security considerations

• Timeouts prevent resource exhaustion on degraded I2P tunnels
• No sensitive information is exposed in error messages
• Error messages are generic enough not to leak timing attacks about key derivation

Implementation details

• Used tokio::time::timeout with appropriate Duration values
• Write operations have shorter timeouts (30s) since they should complete quickly
• Read operations have longer timeouts (60s) to account for I2P network latency
• All error messages are user-facing and descriptive