Repository navigation
chore(vendor): bump tinyagents for the in-memory locator probe fix - #8404
Conversation
Update the vendored tinyagents submodule to a newer commit. Auto-committed-on: cuttlefish Co-authored-by: Medulla <medulla@tinyhumans.ai>
Co-authored-by: Medulla <medulla@tinyhumans.ai>
Record the new chrono dependency for schemars and the schemars dependency for the affected crate in Cargo.lock so the lockfile matches the updated manifests. Auto-committed-on: cuttlefish Co-authored-by: Medulla <medulla@tinyhumans.ai>
A new prefix_reset flag was added to TranscriptMeta, so every construction site now sets it to false to keep the struct literals compiling. Auto-committed-on: cuttlefish Co-authored-by: Medulla <medulla@tinyhumans.ai>
Update test fixtures and call sites to populate the newly added prefix_is_frozen and merge_extra_metadata fields, keeping the existing tests compiling against the extended structs. Auto-committed-on: cuttlefish Co-authored-by: Medulla <medulla@tinyhumans.ai>
Tiny Sweeper reviewVerdict: Incomplete · Priority: none Diagnostics and supporting detailsFull summaryPull request purpose (title): chore(vendor): bump tinyagents for the in-memory locator probe fix. This pull request bumps the vendored `tinyagents` dependency (gitlink plus lockfile) to pick up an upstream fix where `InMemoryTranscriptLocator` probing reserved an empty stem for the next generation, breaking `refresh_prompt_on_resume` on the in-memory store when the system prompt changed between turns. The pin was far behind, so the bump brings three additive struct fields that OpenHuman adapts minimally with values chosen to preserve current behavior. The stated purpose and the full change set: `ThreadMessagePatch::merge_extra_metadata` set to `None` at one production site and two test sites, `TranscriptMeta::prefix_reset: false` across metadata literals in production and test code, and `SessionStateView::prefix_is_frozen` derived as `!prefix.messages().is_empty()` in a single test helper. Unreviewed paths
Partial review reasons
Review snapshot
Completeness: Incomplete What changedThe vendored `tinyagents` gitlink and `Cargo.lock` (gaining `chrono` and `schemars`) are updated, and the resulting new struct fields are filled in at their call sites. Production code gains `merge_extra_metadata: None` in the thread message-update path and `prefix_reset: false` in transcript metadata written by the session host and subagent host. Test fixtures and helpers gain `prefix_reset: false` in metadata literals and `merge_extra_metadata: None` in message patches, and one prompt-refresh test helper gains `prefix_is_frozen: !prefix.messages().is_empty()`. No public API or behavior change is claimed, and the adaptations are neutral values per the description lane. Features
TestsNo supported feature-to-test mapping was produced. Test execution is not inferred. FindingsNo active actionable findings. Before merge
Agent review detailscritique
security
tests
commits
description
Evidence and run details
|
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
No actionable comments were generated in the recent review.
|
| Layer / File(s) | Summary |
|---|---|
Transcript metadata defaults crates/openhuman-core/src/agent/session_host/codec_adoption_tests.rs, crates/openhuman-core/src/agent/session_host/runtime_adapter_tests.rs, crates/openhuman-core/src/agent/session_host/runtime_session/identity.rs, crates/openhuman-core/src/agent/subagent_host/ops/graph/transcript.rs, crates/openhuman-core/src/agent/subagent_host/ops/graph_failed_run_tests.rs, crates/openhuman-core/src/agent/subagent_host/ops/graph_tests.rs, crates/openhuman-core/src/config/migrations/mod_tests.rs, crates/openhuman-core/src/cron/scheduler/origin_delivery_tests.rs, crates/openhuman-core/src/storage/import/sessions_tests.rs, crates/openhuman-core/src/storage/sessions/mod_tests.rs, crates/openhuman-core/src/storage/sessions/splice_fixture_tests.rs, crates/openhuman-core/src/threads/ops/edit_tests.rs, crates/openhuman-core/src/threads/ops/transcript_tests.rs, crates/openhuman-core/src/threads/transcript_host_tests.rs, crates/openhuman-core/src/web_chat/session_checkout_tests.rs |
Transcript metadata fixtures and producers explicitly set prefix_reset to false. |
Restored-prefix state and tinyagents reference crates/openhuman-core/src/agent/session_host/runtime_session/prompt_refresh_tests.rs, vendor/tinyagents |
The resumed-session test helper marks a restored prefix as frozen when it contains messages. The tinyagents submodule reference changes. |
Message patch initialization crates/openhuman-core/src/channels/controllers/ops/connect/memory_tests.rs, crates/openhuman-core/src/threads/ops/crud.rs, crates/openhuman-core/src/threads/store/blocking_tests.rs |
Message patch construction and test inputs explicitly set merge_extra_metadata to None. |
Priority:
Estimated code review effort: 2 (Simple) | ~10 minutes
Change: Bug fix
Possibly related issues
- [chottu-latency] session: resumed threads keep their old prompt when no integrations backend exists #8208 — The resumed-prefix state change and tinyagents submodule update align with the issue’s resume-refresh context.
- tinyhumansai/chottu-core#230 — The PR includes changes related to resumed-session prompt refresh.
- tinyhumansai/chottu-core#224 — The PR includes changes related to resumed-session prompt refresh and prefix state.
Suggested reviewers: m3ga-mind, tinysweeper
Merge Risk: 
Minimal · up to 251b8
-
No actionable merge-blocking risk is identified in the supplied changes.
Pre-merge checks: 
4 · 
1
-
Failed checks: 1 warning Passed checks: 4 passedFull details: Docstring Coverage
-
Explanation
Docstring coverage is 38.10% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 21 functions across 19 files. (1 skipped: 1 unsupported.)
Resolution
Write docstrings for the functions missing them to satisfy the coverage threshold.
-
- Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
A rabbit checks the transcript lines,
Sets prefix flags to clear designs.
Restored messages hold their place,
Patch fields settle into space.
The tinyagents hop completes the chase.
Comment @coderabbitai help to get the list of available commands.
Summary
Bumps
vendor/tinyagentstob23ebadfor tinyhumansai/tinyagents#576. That PR fixesInMemoryTranscriptLocator: probing a session (head_generation,session_exists) registered an empty stem for the next generation, so a followingbegin_generation_from_baselinefailed with "session generation already exists or is reserved". This hitrefresh_prompt_on_resumeon the in-memory store (tests, Mongo-less deployments) whenever the system prompt changed between turns.The pin was 102 tinyagents commits behind, so the bump also brings three additive struct fields. OpenHuman is adapted minimally, with values chosen to keep current behaviour:
ThreadMessagePatch::merge_extra_metadata:None(no JSON merge patch; the existingextra_metadatareplacement is unchanged). 1 production site and 2 test sites.TranscriptMeta::prefix_reset:false(the default; successors keep the legacy prompt-boundary recovery). 2 production sites and 15 test sites.SessionStateView::prefix_is_frozen(test helper only):!prefix.messages().is_empty(), the same derivation the tinyagents runtime uses. OpenHuman's hooks read onlyresumedandhistory, so production behaviour is unchanged.The gitlink bump and the field adaptations are separate commits. The lockfile gains
chronoandschemarsfrom the tinyagents side. This overlaps the tinyagents bump in #8397, so that epic simply merges main.Changes to public API or behavior
None.
Validation
cargo check --workspace --all-targets --features "$(bash scripts/ci/product-features.sh --core)": clean.cargo nextest run -p openhuman --lib --features "<core>": 10441 passed, 1 failed, 4 skipped. The failure,modules::channel_runtime::tests::released_runtime_scopes_callbacks_and_replays_one_final_send, panics withverified artifact path: NotPresent(a prebuilt module artifact is missing locally) and is unrelated to this change.agent::,threads::,attachments::,channels::,cron::under nextest: 2052/2052 passed.cargo clippy -p openhuman --lib --tests -- -D warningsreports only warnings in files this PR does not touch (channels/routes_tests.rs,integrations/composio/ops/scope_prefs.rs,profiles/lifecycle_tests.rs,channels/host/channel_events.rs).probing_for_the_head_generation_does_not_reserve_the_next_one; OpenHuman has no seam that drives the in-memory locator throughrefresh_prompt_on_resumeend to end.Summary by CodeRabbit