Decide the archaic asset staging split; both modes were disagreeing silently

[JamesKane]

archaic_* appeared in neither PATTERNS nor ON_DEMAND_PREFIXES.

Those two lists are the same decision written once per source mode — local-dir mode bundles what PATTERNS names, release mode bundles everything the manifest names except ON_DEMAND_PREFIXES. So an asset missing from both is bundled in CI and absent from dev builds. Nobody chose that: the archaic design's M1 said to check Asset 1's size against the on-demand policy "before deciding", and no decision was ever recorded.

Measured cost on the release it shipped in. All five archaic assets (105.3 MB) went into v0.1.0-alpha.14:

Installer	alpha.13	alpha.14
macOS .dmg	136 MB	199 MB
Windows .exe	105 MB	163 MB
Linux .AppImage	125 MB	186 MB

The split

Same principle as ancestry_haps_ — bundle what users can reach, download what almost nobody does.

Assets	Size	Decision	Why
Tier A archaic_markers_, archaic_marker_dist_	33 MB	Bundled	The marker count runs on chip data as well as WGS, so every user can reach it and it must work with no network. This is also what M1 assumed.
Tier B archaic_outgroup_af_, archaic_classify_, archaic_callable_	72 MB	On demand	Segment calling needs whole-genome data and an hours-long de-novo calling pass to have already completed — a very small reachable population for two thirds of the archaic payload.

Verification

Not assumed — checked:

• Simulated both modes against the real manifest. Each now stages an identical 13 files / 312.2 MB, with divergence empty in both directions. Before: release 384.2 MB vs local 312.2 MB.
• Ran the script. 14 files staged (13 + manifest); Tier A present, Tier B and ancestry_haps absent.
• The download path already exists and is on the only route that loads them. call_archaic_segments_for_subject calls ensure_ancestry_asset for all three before reading (haplogroup.rs:3596). They're in the manifest, so they're sha256-verified on fetch like every other asset.
• First-run seeding is a directory scan (seed_assets_from), so an unbundled asset is simply not seeded rather than erroring.
• bash -n clean.

Expected effect: ~40 MB off each installer (72 MB raw, compressed), taking the .dmg back to roughly 158 MB.

Also

Documents the trap in packaging-and-release.md, because the next asset added will hit it: silence is not a default, it is the CI-only bundle. The asset table gains explicit rows for both archaic tiers.

Companion to #37, which records the same finding in the archaic design doc's open-follow-ups list.

🤖 Generated with Claude Code