Add Electrum connection timeout to prevent thread pool exhaustion

[ben-kaufman]

Summary

Bumps the Synonym Fork to v0.7.0-rc.33 and adds a connection_timeout_secs field to ElectrumSyncConfig (default: 10 s). This bounds Electrum socket operations for both the BDK on-chain and LDK tx-sync clients, preventing Tokio's blocking thread pool from being exhausted by threads stuck on dead sockets under total packet loss (e.g. a captive portal on iOS).

Changes

• Added connection_timeout_secs: u64 field to ElectrumSyncConfig (default: 10 s). Set to 0 to disable; values above 255 are capped to 255 with a warning logged.

Breaking change: existing struct-literal construction of ElectrumSyncConfig must add the new connection_timeout_secs field or switch to ElectrumSyncConfig { .., ..Default::default() }.

Testing

• Verify Electrum sync succeeds under normal network conditions.
• Verify that connecting to a dead/unreachable Electrum server times out within ~10 seconds rather than hanging indefinitely.
• Verify that setting connection_timeout_secs = 0 disables the timeout.
• Verify that values above 255 are capped and a warning is logged.
• Run existing CI test suite across Rust, Kotlin, Swift, and Python bindings.

[claude]

Code review

One issue found: missing blank line between test functions (cargo fmt violation).

[claude]

Missing blank line between test functions — cargo fmt would insert one here.

Every other top-level item boundary in this file uses a blank line separator (47 occurrences of }\n\n#[ vs. this single }\n#[). cargo fmt enforces blank lines between top-level items.

See CLAUDE.md rule: "ALWAYS run cargo fmt before committing to ensure consistent code formatting".

Suggested change

	}
	#[tokio::test(flavor = "multi_thread", worker_threads = 1)]
	}
	#[tokio::test(flavor = "multi_thread", worker_threads = 1)]