v0.3.2

[0.3.2] - 2026-06-19

Fixed

• HLS/DASH playback survives transient network loss. Previously a
  single fetch failure logged a warning and bumped next_play_seq,
  so a brief outage burned through the segment list in seconds and
  silently killed the stream. The player now:
  • Retries the same seq with exponential backoff (500 ms → 1 s →
    2 s → 4 s, capped at 5 s) and transitions to Buffering while
    reconnecting, flipping back to Playing on the first success.
  • Detects the live-edge-rolled-past case: when the requested
    seq has fallen off the front of the live window (typical after
    a long outage), the fetcher returns SegFetch::SkipAhead(front)
    and the loop resyncs to the current window front instead of
    spinning forever on an expired seq.
  • Logs reconnect attempts (retry #N in Xms) and a recovery line
    (reconnected after N failed attempt(s)) so outages are visible
    in journalctl without a debug build.

Full Changelog: v0.3.1...v0.3.2