fix: ignore library keepalive pings when asserting mock wire traffic

[KinseyD]

Summary

The post-merge CI run on main (36599413518) failed on windows-latest: two dynamic_channels integration tests asserted exact wire traffic and intermittently saw an extra PING <timestamp> line.

Root cause

The irc crate's Pinger builds its timer with tokio::time::interval, whose first tick fires immediately. The pinger is enabled on RPL_ENDOFMOTD, so the client sends one PING <Local::now().timestamp()> right after registration instead of after 180s. When that write lands on the wire depends on worker-thread scheduling: on a loaded CI runner it can be delayed into the wire() fixture's capture window, breaking strict equality assertions. Same tree passed the PR run and failed the main run — pure timing flake, no production impact.

Fix

wire() in tests/dynamic_channels.rs now skips PING lines whose payload is pure ASCII digits (the library keepalive). The PING :wire-barrier sentinel carries a non-numeric payload and is unaffected. No production code changes.

Verification

• 50 consecutive dynamic_channels runs green, including 8 rebuild-then-first-run cycles reproducing the CI cold-start condition.
• Full suite 329 passed / 0 failed on Windows; cargo fmt --all --check and cargo clippy --all-targets -- -D warnings clean.