Skip to content

feat: expose native APIs for host memory migrations - #38

Merged
senamakel merged 22 commits into
epic/8000-memory-to-tinymemoryfrom
epic-8000-native-recall-settings-hash
Oct 19, 2026
Merged

senamakel merged 22 commits into
epic/8000-memory-to-tinymemoryfrom
epic-8000-native-recall-settings-hash

Conversation

@senamakel

@senamakel senamakel commented Oct 19, 2026 •

Copy link
Copy Markdown
Owner

Summary

Provide the native APIs required by the remaining OpenHuman memory migrations:

  • Derive Hash for RecallSettings to fingerprint native settings.
  • Expose TurnCache::is_idle with the existing prefetch semantics.
  • Add the cloneable RunStartGate for atomic peer checks across independent background-driver registries.
  • Make the existing installed-host-engine accessor public, without constructing, caching, or rewrapping an engine.

All four prerequisite branches and their individual commits are preserved through ordinary merges. This branch includes the merged Cortex digest and live-harness fix from PR #36.

Refs senamakel/openhuman#91, senamakel/openhuman#92, senamakel/openhuman#94, senamakel/openhuman#95
Part of tinyhumansai/openhuman#8000

Validation status

Focused validation passed on aggregate head 8c28edf30300f003a7449c3592de4021af42b170. All 31 local gates and all 34 hosted checks passed. Coverage also passed on the same aggregate head.

The focused suite passed all 15 commands: 32 exact contract tests in each feature mode, strict Clippy for both affected crates in both modes, and workspace formatting. Full native validation passed on this same head: 31 gates across 32 commands, including 61 feature combinations, strict workspace Clippy, docs, examples, API dependency boundaries, supply-chain checks, and the minimum Rust version build. Workspace tests passed 1,739 all-features and 1,194 default executions with zero failures or ignored tests. Both hosted Cortex live jobs passed 21 actual-service tests each. The coverage run passed 1,718 executions using exactly 25 executed objects and fresh profiles. Official LLVM line coverage is 21,388/22,982 (93.06%); unique LCOV line coverage is 20,180/21,456 (94.05%). All 14 changed executable lines are covered, including every mapped added line in each affected production file. The Hash derive has zero added executable lines. No foreign source records or LLVM warnings were present. Earlier individual-branch results are retained as historical evidence and are not claimed as aggregate validation.

The individual additions were developed with compile regressions for the missing APIs, then focused behavior tests. Added coverage exercises all eleven recall fields, active/queued/drained and invalidated cache work, independent driver status with a shared gate, panic recovery, and installed-engine replacement/clearing with retained Arc identity.

…tinymemory-tools/src/run_slots_

Auto-committed-on: mac-mini
…tinymemory-tools/src/run_slots_

Auto-committed-on: mac-mini
…makel/work/workflow-openhuman/worktrees/epic-8000-host-92-turn-cache/openhuman/vendor/tinymemory into epic-8000-native-host-prerequisites
…makel/work/workflow-openhuman/worktrees/epic-8000-host-94-backfill-import/openhuman/vendor/tinymemory into epic-8000-native-host-prerequisites
@senamakel senamakel changed the title feat(tools): make recall settings hashable feat: expose native APIs for host memory migrations Oct 19, 2026
@senamakel
senamakel merged commit 8c9a2e1 into epic/8000-memory-to-tinymemory Oct 19, 2026
34 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant