Skip to content

Pin validated OmnipodKit backports for early-enable experiment - #1

Open
manudrijvers wants to merge 1 commit into
trio-o5-dex-directfrom
o5-validated-priority-backports
Open

manudrijvers wants to merge 1 commit into
trio-o5-dex-directfrom
o5-validated-priority-backports

Conversation

@manudrijvers

Copy link
Copy Markdown
Owner

Pins OmnipodKit to 654f93a652535940811dc2266c9546ddd213c2d3, the exact validated handshake/key-decoding and faulted-bolus backport candidate in OmnipodKit PR #2. This makes that candidate available on trio-o5-dex-direct for a signed Mac build. The complete Trio diff is the single OmnipodKit gitlink; application source, build configuration and the existing post-insertion control-only hook are unchanged.

Merge OmnipodKit PR nightscout#2 first using a merge commit so this exact pin remains in its history. A squash or rebase requires updating and reverifying this pin before merging here.

Validation:

  • Verified the committed Kit's complete changed-file set and hashes against the retained candidate manifest; packaged checksums and preserved activation/control code also match. Kit tree: 3e1c22f5e7e2488980966fa842d02042088057c0.
  • Repeated portable suite: 127 tests, one existing skip, zero failures. Repeated source-extracted handshake suite: eight scenarios, zero failures. Swift 6.2.4 in the pinned, network-disabled container.
  • The exact candidate already passed 50 native tests and the complete unsigned arm64 Trio device build on November 12, retained in research PR #10. This promotion reuses those results through source equivalence; no new Mac build is claimed.
  • Trio changes only the Kit pin; both checkout states and whitespace checks passed.

Trio candidate commit: 64fd603939a45fab7cb214e94f356cbfb61ecb9e.
Expected journal runnerSourceSHA256: d74e804f70fdb386d46634b05b03ae136fc9690af31fb1f50d4c96b878908442.

After merge, the remaining installation step is a signed build on the owner's Mac using the normal Trio signing process. Preserve the validated device-build settings (-enableThreadSanitizer NO, ENABLE_THREAD_SANITIZER=NO, ENABLE_PREVIEWS=NO) and the frozen formatter's CI=true guard so it does not rewrite sources. The successful validation used Debug with the complete Trio scheme. Verify the source checksums before and after building and record the installed revisions.

Before a physical trial, confirm that Pod Diagnostics → Pump Manager Details exposes the activation journal, the expected source stamp and build-generated Trio/Kit revisions. The experimental option remains off by default and is selected before pairing a fresh Pod; its conditional one-shot enable happens only after insertion/second-prime completion. Sensor pairing remains a separate experiment. This PR does not install an application or establish a real-Pod, delivery-safety or E2-prevention result.

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