feat(pathfinding): .swb format v8 compact index (#3b) (#2471)

## Summary
Phase #3b (final roadmap item), stacked on #2470. Compacts the index trailer from 20 to 8 bytes/chunk.

Trammel: 19.2 MB → 17.9 MB. Roadmap total: 565 MB → 17.9 MB (−96.8%).

## Details
- Trailer stores `{ u32 packedKey = (ChunkX << 16) | ChunkY, u32 recordLength }` per chunk, in record write order; the file offset is dropped and reconstructed by cumulative recordLength from HeaderSize.
- No record reordering, no varint; fixed-stride, TryReadChunk unchanged.
- Also simplifies the accumulated `.swb` code comments across the stack.
- Format v8; v7 files rejected and re-baked once.

## Tests
v8 multi-chunk round-trip (cumulative offset reconstruction) + the v6/v7 suite; full pathfinding suite green; Release build clean.
This commit is contained in:
Kamron Batman 2026-06-07 01:22:20 -07:00 • committed by GitHub
parent c265bcbb5e
commit c7aaf33de9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
16 changed files with 960 additions and 868 deletions

View file

@ -7,9 +7,7 @@ namespace Server.Tests.Pathfinding;
// v7 = per-chunk libdeflate compression on top of the v6 predictive-Z format. Each record is
// compressed independently (random access preserved) behind a u32 uncompressed-length prefix;
// records that do not shrink (tiny Uniform records) are stored raw. StepCacheFileV6Tests already
// round-trips every chunk shape through this compression path; these tests cover the v7-specific
// behavior: that compression engages, that the stored-raw fallback round-trips, and v6 rejection.
// records that do not shrink (tiny Uniform records) are stored raw.
[Collection("Sequential Pathfinding Tests")]
public class StepCacheFileV7Tests
{