brainy/tests/unit/storage/blobStorage.test.ts
David Snelling a3467e1f9b feat: temporal VFS — file content joins the Model-B immutability model
The temporal model had a hole exactly where files were concerned: every
entity write is an immutable generation with before-images, but VFS content
BYTES lived under an eager refCount GC left over from the pre-8.0 design —
unlink could physically destroy bytes that in-window history still
referenced, and overwrite never released the old hash at all (an unbounded
silent leak whose accidental byproduct was the only thing "preserving"
history). Reading the past could therefore return a stale field, a dangling
hash, or nothing, depending on luck.

Fix: blob reclamation becomes a HISTORY decision instead of a LIVENESS
decision. Each blob's metadata now carries historyRefCount alongside the
live refCount:

- The commit seam counts one history reference per persisted before-image
  record carrying a content hash (commitTransaction staging and the
  group-commit flush), recorded BEFORE the record-set persists and carried
  in the generation delta (blobHashes — always present on new deltas, so
  compaction only falls back to reading records for pre-contract
  generations). An aborted transaction compensates best-effort.
- unlink/rmdir/overwrite drop ONLY the live reference (BlobStorage.delete →
  release; overwrite finally releases the superseded hash — cancelling the
  dedup increment on same-content rewrites and closing the leak), and only
  AFTER the canonical mutation commits, so a failed delete can never leave a
  live file whose bytes compaction might reclaim.
- History compaction is the ONE reclamation point: after deleting a
  generation's record-set it releases that set's references and physically
  reclaims any hash at zero live AND zero history references. Pins are
  exempt automatically. Crash ordering is over-count-only in every path
  (record before persist, release after delete), so a crash can leak until
  the scrub recounts but can never reclaim bytes a retained generation
  needs. scrubBlobHistoryRefCounts() restores exactness; existing stores get
  a one-time marker-gated backfill on open, failing into leak-safe mode
  (reclamation disabled) rather than guessing.

On top of the protected history, the temporal API the generational model
always implied:

- vfs.readFile(path, { asOf }) — the exact bytes as of a generation or Date,
  materialized from the history (pinned view released so compaction is
  never blocked by a read).
- vfs.history(path) — FileVersion[] ascending ({ generation, timestamp,
  hash, size, mimeType? }), the newest entry being the live state.
- Overwrites now refresh the file entity's data/embedding text — semantic
  search and the data field previously served the FIRST version's text
  forever (the stale-field defect a consumer's incident recovery depended
  on by luck).

Integration suite (temporal-vfs.test.ts): per-version exact reads +
history listing, leak-fix + history protection on overwrite, rm keeps bytes
readable, compaction reclaims past-window bytes and preserves in-window
(including the cross-file dedup case where an old file's history and a
newer file's removal share one hash), data freshness, and scrub exactness.
2026-07-10 16:43:48 -07:00

526 lines
19 KiB
TypeScript

/**
* Comprehensive tests for BlobStorage (src/storage/blobStorage.ts)
*
* Tests:
* - Content-addressable storage (SHA-256, integrity verification)
* - Deduplication via reference counting
* - Compression (zstd, MIME-aware auto mode)
* - LRU caching (bounded cache still serves correct bytes)
* - Reference counting + delete-at-zero
* - Error handling
* - Performance characteristics
* - Wrapped-binary-data regressions (key-based dispatch)
*
* Cache-bypass pattern: the store has no cache-introspection API (the LRU is
* an internal optimization), so tests that must force a storage read create a
* FRESH BlobStorage over the same adapter — a cold cache by construction.
*/
import { describe, it, expect, beforeEach } from 'vitest'
import { BlobStorage, BlobStoreAdapter } from '../../../src/storage/blobStorage.js'
/**
* Simple in-memory blob store adapter for testing
*/
class InMemoryBlobAdapter implements BlobStoreAdapter {
private store = new Map<string, Buffer>()
async get(key: string): Promise<Buffer | undefined> {
return this.store.get(key)
}
async put(key: string, data: Buffer): Promise<void> {
this.store.set(key, data)
}
async delete(key: string): Promise<void> {
this.store.delete(key)
}
async list(prefix: string): Promise<string[]> {
const keys: string[] = []
for (const key of this.store.keys()) {
if (key.startsWith(prefix)) {
keys.push(key)
}
}
return keys.sort()
}
}
describe('BlobStorage', () => {
let adapter: BlobStoreAdapter
let blobStorage: BlobStorage
beforeEach(() => {
adapter = new InMemoryBlobAdapter()
blobStorage = new BlobStorage(adapter)
})
describe('Content-Addressable Storage', () => {
it('should compute SHA-256 hash of data', () => {
const data = Buffer.from('hello world')
const hash = BlobStorage.hash(data)
expect(hash).toBe('b94d27b9934d3e08a52e52d7da7dabfac484efe37a5380ee9088f7ace2efcde9')
expect(hash).toHaveLength(64)
})
it('should write blob and return hash', async () => {
const data = Buffer.from('test data')
const hash = await blobStorage.write(data)
expect(hash).toBeTruthy()
expect(hash).toHaveLength(64)
})
it('should read blob by hash', async () => {
const data = Buffer.from('test data')
const hash = await blobStorage.write(data)
const retrieved = await blobStorage.read(hash)
expect(retrieved.toString()).toBe('test data')
})
it('should verify data integrity on read', async () => {
const data = Buffer.from('test data')
const hash = await blobStorage.write(data)
// Corrupt the blob data behind the store's back
await adapter.put(`blob:${hash}`, Buffer.from('corrupted'))
// A fresh instance has a cold cache, so read() must hit storage and
// detect the corruption via hash verification.
const coldStore = new BlobStorage(adapter)
await expect(coldStore.read(hash)).rejects.toThrow('integrity check failed')
})
it('should check if blob exists', async () => {
const data = Buffer.from('test data')
const hash = await blobStorage.write(data)
expect(await blobStorage.has(hash)).toBe(true)
expect(await blobStorage.has('0'.repeat(64))).toBe(false)
})
})
describe('Deduplication', () => {
it('should deduplicate identical blobs', async () => {
const data = Buffer.from('duplicate data')
const hash1 = await blobStorage.write(data)
const hash2 = await blobStorage.write(data)
expect(hash1).toBe(hash2)
// Exactly one stored blob + one metadata record
expect(await adapter.list('blob:')).toHaveLength(1)
expect(await adapter.list('blob-meta:')).toHaveLength(1)
})
it('should increment ref count on duplicate write', async () => {
const data = Buffer.from('test data')
await blobStorage.write(data)
const hash = await blobStorage.write(data)
const metadata = await blobStorage.getMetadata(hash)
expect(metadata?.refCount).toBe(2)
})
it('should track every duplicate write in the reference count', async () => {
const data = Buffer.from('x'.repeat(1000))
const hash1 = await blobStorage.write(data)
const hash2 = await blobStorage.write(data)
const hash3 = await blobStorage.write(data)
expect(hash2).toBe(hash1)
expect(hash3).toBe(hash1)
const metadata = await blobStorage.getMetadata(hash1)
expect(metadata?.refCount).toBe(3)
})
})
describe('Compression', () => {
it('should compress large text data with zstd', async () => {
const data = Buffer.from('a'.repeat(10000))
const hash = await blobStorage.write(data, { compression: 'zstd' })
const metadata = await blobStorage.getMetadata(hash)
// zstd may not be available in test environment - falls back to 'none'
// This is expected behavior (see BlobStorage.ensureCompressionReady fallback)
if (metadata?.compression === 'zstd') {
expect(metadata.compressedSize).toBeLessThan(metadata.size)
} else {
// Fallback to 'none' is acceptable when zstd unavailable
expect(metadata?.compression).toBe('none')
}
})
it('should decompress zstd data on read', async () => {
const originalData = Buffer.from('test data '.repeat(100))
const hash = await blobStorage.write(originalData, { compression: 'zstd' })
const retrieved = await blobStorage.read(hash)
expect(retrieved.toString()).toBe(originalData.toString())
})
it('should not compress small blobs', async () => {
const data = Buffer.from('small')
const hash = await blobStorage.write(data, { compression: 'auto' })
const metadata = await blobStorage.getMetadata(hash)
expect(metadata?.compression).toBe('none')
})
it('should skip compression for already-compressed MIME types in auto mode', async () => {
const data = Buffer.from('x'.repeat(5000))
const hash = await blobStorage.write(data, {
compression: 'auto',
mimeType: 'image/jpeg'
})
const metadata = await blobStorage.getMetadata(hash)
expect(metadata?.compression).toBe('none')
})
it('should auto-compress large compressible payloads', async () => {
const data = Buffer.from('x'.repeat(5000))
const hash = await blobStorage.write(data, { compression: 'auto' })
const metadata = await blobStorage.getMetadata(hash)
// Should compress if zstd is available
if (metadata?.compression === 'zstd') {
expect(metadata.compressedSize).toBeLessThan(metadata.size)
}
})
})
describe('LRU Caching', () => {
it('should serve identical bytes from cache and from storage', async () => {
const data = Buffer.from('cached data')
const hash = await blobStorage.write(data)
// Warm read (write-through cache) and cold read (fresh instance)
const warm = await blobStorage.read(hash)
const cold = await new BlobStorage(adapter).read(hash)
expect(warm.equals(data)).toBe(true)
expect(cold.equals(data)).toBe(true)
})
it('should keep serving correct bytes when the cache evicts under pressure', async () => {
const smallCache = new BlobStorage(adapter, { cacheMaxSize: 100 })
// Write blobs that exceed cache size (forces LRU eviction)
const blobs = [
Buffer.from('x'.repeat(50)),
Buffer.from('y'.repeat(50)),
Buffer.from('z'.repeat(50))
]
const hashes: string[] = []
for (const blob of blobs) {
hashes.push(await smallCache.write(blob))
}
// Every blob still reads back correctly, evicted or not
for (let i = 0; i < blobs.length; i++) {
const retrieved = await smallCache.read(hashes[i])
expect(retrieved.equals(blobs[i])).toBe(true)
}
})
})
describe('Reference Counting', () => {
it('should track reference count', async () => {
const data = Buffer.from('test')
const hash = await blobStorage.write(data)
let metadata = await blobStorage.getMetadata(hash)
expect(metadata?.refCount).toBe(1)
// Write duplicate (increments ref count)
await blobStorage.write(data)
metadata = await blobStorage.getMetadata(hash)
expect(metadata?.refCount).toBe(2)
})
it('release() drops live references; bytes are reclaimed ONLY by reclaimIfUnreferenced at zero-zero', async () => {
const data = Buffer.from('test')
// Write twice (live refCount = 2)
const hash = await blobStorage.write(data)
await blobStorage.write(data)
// Release once (refCount = 1, blob still exists)
await blobStorage.release(hash)
expect(await blobStorage.has(hash)).toBe(true)
// Release again (refCount = 0) — bytes STILL exist: content is
// immutable under the temporal model; only history compaction reclaims.
await blobStorage.release(hash)
expect(await blobStorage.has(hash)).toBe(true)
expect((await blobStorage.getMetadata(hash))?.refCount).toBe(0)
// A history reference blocks reclamation even at live-zero.
await blobStorage.recordHistoryReference(hash)
expect(await blobStorage.reclaimIfUnreferenced(hash)).toBe(false)
expect(await blobStorage.has(hash)).toBe(true)
// Drop the history reference → zero-zero → reclaim succeeds.
await blobStorage.releaseHistoryReference(hash)
expect(await blobStorage.reclaimIfUnreferenced(hash)).toBe(true)
expect(await blobStorage.has(hash)).toBe(false)
})
})
describe('Metadata', () => {
it('should store blob metadata', async () => {
const data = Buffer.from('test')
const hash = await blobStorage.write(data, { compression: 'none' })
const metadata = await blobStorage.getMetadata(hash)
expect(metadata?.hash).toBe(hash)
expect(metadata?.size).toBe(data.length)
expect(metadata?.compression).toBe('none')
expect(metadata?.createdAt).toBeGreaterThan(0)
expect(metadata?.refCount).toBe(1)
})
it('should return undefined metadata for unknown hashes', async () => {
expect(await blobStorage.getMetadata('f'.repeat(64))).toBeUndefined()
})
})
describe('Error Handling', () => {
it('should throw on reading non-existent blob', async () => {
await expect(
blobStorage.read('f'.repeat(64))
).rejects.toThrow('Blob metadata not found')
})
it('should throw on reading blob with missing metadata', async () => {
const data = Buffer.from('test')
const hash = await blobStorage.write(data)
// Delete metadata but keep blob bytes
await adapter.delete(`blob-meta:${hash}`)
// A fresh instance has a cold cache, so read() must consult metadata
const coldStore = new BlobStorage(adapter)
await expect(coldStore.read(hash)).rejects.toThrow('metadata not found')
})
})
describe('Performance', () => {
it('should write 1000 small blobs quickly', async () => {
const start = Date.now()
for (let i = 0; i < 1000; i++) {
await blobStorage.write(Buffer.from(`blob${i}`))
}
const elapsed = Date.now() - start
expect(elapsed).toBeLessThan(5000) // Should complete in < 5s
})
it('should read 1000 cached blobs very quickly', async () => {
// Write and cache blobs
const hashes: string[] = []
for (let i = 0; i < 1000; i++) {
const hash = await blobStorage.write(Buffer.from(`blob${i}`))
hashes.push(hash)
}
// First read (populate cache)
for (const hash of hashes) {
await blobStorage.read(hash)
}
// Second read (from cache)
const start = Date.now()
for (const hash of hashes) {
await blobStorage.read(hash)
}
const elapsed = Date.now() - start
expect(elapsed).toBeLessThan(1000) // Should be very fast from cache
})
})
describe('Wrapped-binary regression (v5.10.0) - hash verification on unwrapped bytes', () => {
it('should handle wrapped binary data without hash mismatch', async () => {
// Import TestWrappingAdapter that actually wraps data like production
const { TestWrappingAdapter } = await import('../../helpers/TestWrappingAdapter.js')
const wrappingAdapter = new TestWrappingAdapter()
const testBlobStorage = new BlobStorage(wrappingAdapter)
const originalData = Buffer.from('test content for v5.10.0 regression test')
// Write blob (TestWrappingAdapter wraps as {_binary: true, data: "base64..."})
const hash = await testBlobStorage.write(originalData)
// Fresh instance = cold cache: read() must fetch + unwrap from storage.
// v5.10.0 bug: read() hashed the wrapped bytes instead of the unwrapped
// content; the fix runs unwrapBinaryData() before hash verification.
const retrieved = await new BlobStorage(wrappingAdapter).read(hash)
expect(retrieved.equals(originalData)).toBe(true)
expect(retrieved.toString()).toBe('test content for v5.10.0 regression test')
})
it('should handle multiple wrapped blobs without integrity errors', async () => {
const { TestWrappingAdapter } = await import('../../helpers/TestWrappingAdapter.js')
const wrappingAdapter = new TestWrappingAdapter()
const testBlobStorage = new BlobStorage(wrappingAdapter)
const testData = [
Buffer.from('first blob'),
Buffer.from('second blob'),
Buffer.from('third blob with more content'),
Buffer.from(JSON.stringify({ test: 'json data' })),
]
const hashes: string[] = []
// Write all blobs
for (const data of testData) {
const hash = await testBlobStorage.write(data)
hashes.push(hash)
}
// Cold-cache instance forces storage reads
const coldStore = new BlobStorage(wrappingAdapter)
for (let i = 0; i < hashes.length; i++) {
const retrieved = await coldStore.read(hashes[i])
expect(retrieved.equals(testData[i])).toBe(true)
}
})
it('should work even if adapter fails to unwrap (defense-in-depth)', async () => {
// Create a buggy adapter that doesn't unwrap properly
class BuggyAdapter implements BlobStoreAdapter {
private storage = new Map<string, any>()
async get(key: string): Promise<any | undefined> {
// Simulate a bug where adapter returns wrapped data instead of Buffer
const data = this.storage.get(key)
if (!data) return undefined
// Return wrapped object (bug) instead of unwrapped Buffer
return data // {_binary: true, data: "base64..."}
}
async put(key: string, data: Buffer): Promise<void> {
// Store as wrapped object (like real storage)
this.storage.set(key, {
_binary: true,
data: data.toString('base64')
})
}
async delete(key: string): Promise<void> {
this.storage.delete(key)
}
async list(prefix: string): Promise<string[]> {
return Array.from(this.storage.keys()).filter(k => k.startsWith(prefix))
}
}
const buggyAdapter = new BuggyAdapter()
const testBlobStorage = new BlobStorage(buggyAdapter)
const originalData = Buffer.from('test defense-in-depth')
const hash = await testBlobStorage.write(originalData)
// Even with buggy adapter (and a cold cache), read() should unwrap and
// verify correctly
const retrieved = await new BlobStorage(buggyAdapter).read(hash)
expect(retrieved.equals(originalData)).toBe(true)
})
})
describe('Key-based dispatch regression (v6.2.0)', () => {
it('should handle JSON-like compressed data without integrity failures', async () => {
// THE KILLER TEST CASE: Data that looks like JSON when compressed
// This would fail with v5.10.1 wrapBinaryData() guessing approach
const { TestWrappingAdapter } = await import('../../helpers/TestWrappingAdapter.js')
const wrappingAdapter = new TestWrappingAdapter()
const testBlobStorage = new BlobStorage(wrappingAdapter)
// Create JSON data that will be compressed
const jsonData = { nested: { key: 'value', array: [1, 2, 3] }, repeated: 'x'.repeat(1000) }
const originalData = Buffer.from(JSON.stringify(jsonData))
// Write blob (compression happens, might create JSON-parseable bytes)
const hash = await testBlobStorage.write(originalData)
// v5.10.1 bug: If compressed bytes accidentally parse as JSON,
// wrapBinaryData() would store parsed object instead of wrapped binary
// On read: JSON.stringify(object) !== original compressed bytes → hash mismatch
//
// v6.2.0 fix: Key-based dispatch eliminates guessing
// 'blob:hash' → Always wrapped as binary, never parsed
const retrieved = await new BlobStorage(wrappingAdapter).read(hash)
// Hash MUST match
expect(BlobStorage.hash(retrieved)).toBe(hash)
expect(retrieved.equals(originalData)).toBe(true)
})
it('should handle metadata keys correctly', async () => {
// Verify that key-based dispatch correctly handles metadata keys
// This test verifies the dispatch logic, not the full read/write cycle
const { TestWrappingAdapter } = await import('../../helpers/TestWrappingAdapter.js')
const wrappingAdapter = new TestWrappingAdapter()
// Test metadata key dispatch: should parse as JSON
const metadataObj = { hash: 'test123', size: 42, compression: 'none' as const }
const metadataBuffer = Buffer.from(JSON.stringify(metadataObj))
await wrappingAdapter.put('blob-meta:test123', metadataBuffer)
const retrieved = await wrappingAdapter.get('blob-meta:test123')
// Should be parsed as JSON object (not wrapped as binary)
expect(retrieved).toBeDefined()
expect(typeof retrieved).toBe('object')
expect(retrieved.hash).toBe('test123')
expect(retrieved.size).toBe(42)
})
it('should never parse blob bytes as JSON on the write path', async () => {
// Verify that the blob bridge uses key-based dispatch,
// NOT wrapBinaryData() guessing
const { TestWrappingAdapter } = await import('../../helpers/TestWrappingAdapter.js')
const wrappingAdapter = new TestWrappingAdapter()
const testBlobStorage = new BlobStorage(wrappingAdapter)
// Create data that would trigger wrapBinaryData() bug if used
const problematicData = Buffer.from('[{"looks":"like","valid":"json"}]')
const hash = await testBlobStorage.write(problematicData)
const retrieved = await new BlobStorage(wrappingAdapter).read(hash)
// Should retrieve exact same bytes (no JSON parsing occurred)
expect(retrieved.equals(problematicData)).toBe(true)
expect(BlobStorage.hash(retrieved)).toBe(hash)
})
})
})