Skip to content

Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors

Rust

docs(store): finish strict V62 migration guidance

AkitaOnRails6d20b73e9384
diff --git a/crates/ai-memory-store/src/migrations.rs b/crates/ai-memory-store/src/migrations.rs
index 8985b925..e56e62c2 100644
--- a/crates/ai-memory-store/src/migrations.rs
+++ b/crates/ai-memory-store/src/migrations.rs
@@ -1631,6 +1631,16 @@ mod tests {
.checksum();
assert_eq!(checksum, ORIGINAL_V62_CHECKSUM);
+ let embedded = migrations::runner()
+ .get_migrations()
+ .iter()
+ .find(|migration| migration.version() == V62_VERSION)
+ .expect("V62 must remain embedded")
+ .checksum();
+ assert_ne!(
+ embedded, checksum,
+ "the V62 repair is unnecessary if the embedded SQL matches v2.2.1"
+ );
}
/// The one known V62 checksum is repaired before the strict runner sees
diff --git a/docs/design-memory-aging.md b/docs/design-memory-aging.md
index 3d5a6ee5..b164ee5d 100644
--- a/docs/design-memory-aging.md
+++ b/docs/design-memory-aging.md
@@ -428,7 +428,8 @@ divergence check. V62 is the historical exception. Its exact v2.2.1 checksum is
repaired to the embedded checksum before the runner starts (#1443), after which
`set_abort_divergent(true)` keeps every unknown edit fatal. `set_abort_missing(true)`
also keeps a store *ahead* of the binary fail-closed. Future changes must never
-edit a shipped migration in place or broaden the narrow V62 repair.
+edit a shipped migration in place or broaden the narrow V62 repair;
+`scripts/check-migrations-immutable.sh` enforces that rule in CI.
**Any data backfill runs on the boot path, not inside the migration
transaction**: chunked, resumable, and WAL-checkpointed, mirroring the #776
@@ -495,9 +496,9 @@ content.
### 4. Downgrade note
-Additive columns plus `set_abort_missing(true)` (`migrations.rs:34`) mean an older
+Additive columns plus `set_abort_missing(true)` in `migrations::run` mean an older
binary opening a newer store **fails closed** with the actionable
-`StoreError::DataSchemaAhead` (`error.rs:36`, raised at `migrations.rs:55`). It
+`StoreError::DataSchemaAhead` raised by `classify_run_error`. It
names the applied-vs-supported versions and refuses to open, rather than silently
corrupting a store it doesn't understand. That guarantee must be preserved by
every new migration.
diff --git a/docs/design-page-ingestion-windows.md b/docs/design-page-ingestion-windows.md
index 95fd2a40..486768ec 100644
--- a/docs/design-page-ingestion-windows.md
+++ b/docs/design-page-ingestion-windows.md
@@ -22,10 +22,10 @@ invoked from `Store::open` next to the entity-index backfill) that
reproduces the identical end state in bounded batches, checkpoints the
WAL (`TRUNCATE`) between each, logs progress, resumes rather than
restarts if interrupted, and no-ops on a store already backfilled. The
-in-place reshape changed V62's checksum. The migration runner repairs only
-the exact checksum recorded for the SQL shipped in v2.2.1 before running
-strict validation; correctly-migrated stores still open, while every unknown
-checksum divergence remains fatal (#1443).*
+in-place reshape changed V62's checksum. `migrations::run` repairs only the
+exact checksum recorded for the SQL shipped in v2.2.1 before running strict
+validation; correctly-migrated stores still open, while every unknown checksum
+divergence and every store ahead of the binary remains fatal (#1443).*
## 1. The problem, and why entity-link windows are not enough