Write registry ids for sounds, not id + 1

[Pix3lPirat3]

The 1.19.2+ folders added one to every sound id because the packet encodes registry id + 1 (0 = inline). node-minecraft-protocol's registryEntryHolder already hands consumers the registry id and node-minecraft-data indexes sounds by the id field, so mineflayer looked up the neighbouring sound on 1.19.2 - 26.1. Write the registry id in all twelve folders, like the 1.19 and 1.20.4 modules do.

---

Companion PRs: regenerated data is in PrismarineJS/minecraft-data#1316.

[rom1504]

Astra agent review — AI-generated, not manually written by the maintainer.

The producer correction looks sound: raw registry IDs belong in sounds.json; the packet-specific +1 belongs to the registry-holder codec. I verified successful per-version CI jobs at this exact head and compared their emitted sounds.json with #1316 for 1.19.2, 1.20, 1.21.3, 1.21.8 and 26.1: all five match exactly. Production NMP codec → current Mineflayer sound-plugin probes then resolve chicken.hurt correctly across those protocol versions, including 26.1 ID355.

These are specific successful jobs in an overall cancelled matrix, not a claim that the full matrix passed. I found no new code blocker; preserve the companion data publication so existing consumers receive the corrected IDs.

Skills used: prismarine-protocol-data-review separated registry IDs from wire sentinels and checked generated outputs through actual consumers; prismarine-architecture-review placed the normalization in the producer; prismarine-review checked exact-head CI evidence and scope.

[rom1504]

Astra agent review — AI-generated, not manually written by the maintainer.

The sound-registry correction remains correct in the reviewed scope: the producer exports zero-based registry IDs, while the codec owns the wire sentinel offset. The earlier exact-head artifacts and production NMP → Mineflayer probes remain applicable; see the existing review.

The remaining landing condition is integration: the refreshed matrix contains failed Build 1.14, 1.20.4, and 1.21 jobs. Diagnose or rerun those jobs before treating the whole generator PR as ready; successful targeted artifacts do not clear the rest of the matrix. I have no new code defect to repeat, and did not rerun Java generation locally.

Skills used: prismarine-review checked the current revision and discussion; prismarine-protocol-data-review checked the version-selected producer/consumer contract; prismarine-architecture-review checked package ownership and integration scope.