Group-managed datasets must never have Init/Sync/Run called on them.
Rather than patching with NopSyncer, introduce View[T] — a thin wrapper
that exposes only Load(). The compiler now prevents misuse: callers can
read values but cannot drive fetch/reload cycles directly.
Dataset[T] no longer needs a syncer when owned by a Group; View.reload()
delegates to the inner Dataset.reload() for Group.reloadAll().
- gitshallow: replace in-place Depth mutation with effectiveDepth() method;
remove depth normalisation from New() since it was masking the bug
- ipcohort: extract sortNets() helper using cmp.Compare, eliminating 3 identical
sort closures; add ContainsAddr(netip.Addr) for pre-parsed callers; guard
Contains() against IPv6 panic (As4 panics on non-v4); add IPv6 test
- dataset: Add() now sets NopSyncer{} so callers cannot panic by accidentally
calling Init/Sync/Run on a Group-managed Dataset
dataset.Loader[T] is now func() (*T, error) — a closure capturing its own
paths/config, so multi-file cases (LoadFiles(paths...)) work naturally.
Dataset.Close func(*T) is called with the old value after each swap, enabling
resource cleanup (e.g. geoip2.Reader.Close).
Sources.Datasets() builds a dataset.Group + three typed *Dataset[ipcohort.Cohort].
main.go now uses blGroup.Run / cityDS.Run / asnDS.Run instead of hand-rolled
atomic.Pointer + polling loops. containsInbound/OutBound accept *Dataset[Cohort].
nopSyncer handles file-only GeoIP paths (no download, just open).
Dataset[T]: one Syncer + one Loader + one atomic.Pointer. Init/Sync/Run.
Group: one Syncer driving N datasets — single Fetch, all reloads fire
together. Add[T](g, loader, path) registers a typed dataset in the group.
Discovered organically: the reload+atomic-swap pattern repeated across
every cmd is exactly this abstraction.
