net: drop explicit fetch_blob workaround once iroh-docs retries downloads

[fosskar]

Blocked on upstream: n0-computer/iroh-docs#88 (open; 2026-07-01 update added per-(namespace, peer) debounce + per-blob backoff addressing the maintainer's overhead concern). iroh-docs downloads entry content once on arrival and never retries; Node::fetch_blob (called from the WriteFile arm on local miss, 30s bound) is our workaround.

The on-write peer fetch stays correct either way, but becomes dead weight once upstream merges and the minor-pinned iroh stack is bumped.

Refs: TODO.md Open #8, AGENTS.md export-path notes.

[fosskar]

Upstream status (from the retired TODO.md, 2026-07-01): iroh-docs#88 added per-(namespace, peer) debounce + per-blob backoff to address the maintainer's overhead concern - still open, watching; no action until merged and the pin bumps.