@blockcast/mmt-container 1.0.1 → 2.0.0-main.11614d24ad32

package/README.md

@@ -158,11 +158,13 @@ The constructor rejects non-null, non-finite, or non-positive deadlines. Reliabl
 ### CARP / CMSF catalog (`./carp-catalog`)
 
 Validate the published catalog schema through the package API, then run
-`validateCmsfMmtpCatalog`: the multicast rules of draft-ramadan-moq-multicast-01
-that JSON Schema cannot state (a repeated `kid` or `firstEpoch` in
-`multicast.auth.keys`, a `manifestPort` that clashes with a port of its
-`(S, G)`, a track declaring a foreign namespace) live only in the runtime
-validator.
+`validateCmsfMmtpCatalog`. The multicast rules of draft-ramadan-moq-multicast-02
+4.1 / 7.2.6 (a repeated `kid` or `firstEpoch` in `multicast.auth.keys`, two
+endpoints on one `(S, G, port)`, an address that is not an RFC 3986 IP literal,
+a track declaring a foreign namespace) are never a catalog fault: the
+validator does not throw on them, and `admittedMulticastEndpoints` /
+`multicastCatalogIssues` of `@blockcast/mmt-base` name the endpoints a receiver
+ignores (none admitted on a catalog-level fault), so MoQ unicast stays usable.
 
 ```ts
 import schema from "@blockcast/mmt-container/catalog.schema.json" with { type: "json" }
@@ -258,7 +260,7 @@ without one: that entry carries the frequency in kHz (REQUIRED) and,
 optionally, the PLP, service id and SLS URI (draft -01 4.2.2/4.2.3) — the only
 tuning information an ATSC3 receiver has (the SLS behind
 `slsUri` describes the MMTP session, MP table included, per A/331:2026-04 7.2,
-and nothing of this profile, so `manifestPort` and keys come from provisioning), and the schema
+and nothing of this profile, so the flow parameters and keys come from provisioning), and the schema
 cannot enforce it because `networkSource` is optional there (a native join has
 none). `version`, `validUntil` and `enabled` are still emitted and still
 outside `$defs/multicast` (the BLO-3323 freshness model, tracked separately).
@@ -275,7 +277,7 @@ per-commit `0.1.0-main.<sha>` prerelease from `stamp-prerelease.mjs`), and a
 > Superseded by BLO-40665. Every package is now bumped in lockstep by a `v*`
 > tag and the sibling ranges move with it, so the objection above no longer
 > applies -- it was an argument against bumping *one* package. The set is at
-> `1.0.1`; see `RELEASES.md`.
+> `2.0.0`; see `RELEASES.md`.
 
 #### Track-role convention (Phase 14)