Reuse current model context across paginated resume config loading (#56224)

## Why

Resuming a paginated thread rechecks stored history after loading configuration. Unchanged model context can be reused instead of loading it again.

## What changed

Retain the prepared model context across configuration loading and reuse it when the thread ID, rollout path, and history revision still match. Add `ThreadStore::is_model_context_current` to validate revisions, with a fresh read when validation fails or is unavailable.

## Testing

Add coverage for revision invalidation after history is appended or removed. Extend resume tests to cover paginated history, concurrent metadata and lifecycle changes, workspace-root changes, and preservation of saved user messages in model requests.

GitOrigin-RevId: bad2ca0caddb63523264ac530ffb801db3961f2b
This commit is contained in:
Charlie Marsh
2026-10-22 17:15:06 +00:00
committed by copyberry
parent 9cb547349e
commit ffda54f5b8
5 changed files with 207 additions and 21 deletions
+13
View File
@@ -268,6 +268,19 @@ pub trait ThreadStore: Any + Send + Sync {
})
}
/// Checks whether a previously loaded model context still represents the stored history.
///
/// Stores that cannot validate the opaque revision must return false. A false result or an
/// error requires a fresh read. This does not acquire writer ownership;
/// [`Self::resume_thread`] must still validate the snapshot after acquisition.
fn is_model_context_current(
&self,
_params: LoadThreadHistoryParams,
_revision: String,
) -> ThreadStoreFuture<'_, bool> {
Box::pin(async { Ok(false) })
}
/// Freezes source history and model context used to initialize a referenced fork.
///
/// Stores without reference-backed fork support can retain this default implementation.