FetchAndDecryptBlob now returns io.ReadCloser with a hashVerifyReader
that computes the double-SHA-256 on-the-fly during reads. Hash is
verified on Close() after the stream is fully consumed. This avoids
loading entire blobs into memory, which could exceed available RAM.
Addresses review feedback on PR #39.
The --prune flag on 'snapshot create' was accepted but silently did nothing
(TODO stub). This connects it to actually:
1. Purge old snapshots (keeping only the latest) via PurgeSnapshots
2. Remove unreferenced blobs from storage via PruneBlobs
The pruning runs after all snapshots complete successfully, not per-snapshot.
Both operations use --force mode (no interactive confirmation) since --prune
is an explicit opt-in flag.
Moved the prune logic from createNamedSnapshot (per-snapshot) to
CreateSnapshot (after all snapshots), which is the correct location.
Adds regression tests for issue #28 (fixed in PR #33) to prevent
reintroduction of the double-close bug in CompressStream.
Tests cover:
- CompressStream with normal input
- CompressStream with large (512KB) input
- CompressStream with empty input
- CompressData close correctness
Blobs are typically hundreds of megabytes and should not be held in memory.
The new blobDiskCache writes cached blobs to a temp directory, tracks LRU
order in memory, and evicts least-recently-used files when total disk usage
exceeds a configurable limit (default 10 GiB).
Design:
- Blobs written to os.TempDir()/vaultik-blobcache-*/<hash>
- Doubly-linked list for O(1) LRU promotion/eviction
- ReadAt support for reading chunk slices without loading full blob
- Temp directory cleaned up on Close()
- Oversized entries (> maxBytes) silently skipped
Also adds blob_fetch_stub.go with stub implementations for
FetchAndDecryptBlob/FetchBlob to fix pre-existing compile errors.
clawbot