docs+tests: HIGH-1 follow-through — README, stale comments, regressio…

[Bluechip23]

…n tests

Aligns the docs and tests with the M-1 fix (commit a35c4e9... f75ad9a) that deferred commit_pool_ordinal allocation from create-time to threshold-cross-time.

Docs

• README.md: rewrote the mint-formula x definition (line 655) so it reflects the new "ordinal among pools that have CROSSED threshold" semantics, and updated the two anti-spam narrative blocks (lines 133, 561) to clarify that the 1h create cooldown + USD fee defend against registry bloat (not against ordinal inflation — the fix removes that vector entirely).

Stale comments

• factory/src/pool_creation_reply.rs:232: comment claimed the ordinal is "captured at create time" — false post-fix. Replaced with a description of the sentinel-then-overwrite flow.

Regression tests (NEW)

• test_commit_pool_ordinal_advances_on_threshold_cross_not_create: pins the core invariant — Create must not bump COMMIT_POOL_COUNTER; NotifyThresholdCrossed must.
• test_junk_creates_do_not_inflate_ordinal_for_legitimate_crosser: exercises the economic invariant — three junk creates from distinct attacker addresses must leave the counter at 0, and a legitimate pool that crosses afterward must receive ordinal 1 (with the mint amount in the undecayed-base range, confirming x=1 in the formula).

If either test fails in the future, the M-1 attack has regressed.

Test count: 250 → 252 (both new tests passing). Full factory suite passes; workspace compiles clean.

https://claude.ai/code/session_019mxdB2vwswjgipCUq2x4Fu