Skip to content

removed experimental features, added expiry warning - #113

Closed
den-saunders wants to merge 7 commits into
LoopKit:mainfrom
den-saunders:main
Closed

den-saunders wants to merge 7 commits into
LoopKit:mainfrom
den-saunders:main

Conversation

@den-saunders

Copy link
Copy Markdown

No description provided.

@bjorkert

Copy link
Copy Markdown

Hi there,

I noticed you've submitted a pull request to the LoopWorkspace project. It looks like the changes might be personal customizations rather than updates intended for the wider Loop community. This is a common occurrence, so no worries!

If these changes were meant just for your personal use, it would be helpful to close this pull request to keep the main project focused on general updates that benefit all users. Closing the pull request would also alleviate some of the workload on the project's maintainers, who review and manage these contributions.

@den-saunders

Copy link
Copy Markdown
Author

Accidental submission to main. Sorry!

jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 12, 2026
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 12, 2026
…art the watch app

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 13, 2026
…oud urgent errors)

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 13, 2026
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 13, 2026
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
jeremybarnum pushed a commit to jeremybarnum/LoopWorkspace that referenced this pull request Aug 13, 2026
…than linkage

CLAUDE.md held this back deliberately — "goes in deliberately once there is real coverage rather
than reachability tests" — because a gate that only proves a target links is cost without
protection. That condition is met. The watch suite is 33 tests covering timer arming and epoch
scoping through the scheduling seam, the LoopKit#113 wedge truth table, and the three dead-man alerts'
arming and replacement, and every one of them has been sabotage-verified to go red.

Costs ~90 s and a second xcodebuild run, because it is a different platform.

Sabotage-verified as a GATE, not just as tests: inverting the hand-back-before-loop-watchdog
assertion makes the script print the assertion, refuse to archive, and exit 68. Confirmed
unpiped — reading $? through a pipe returns tail's status, which is exactly how a failed upload
once got reported as a successful ship.

Failure handling mirrors the iOS gate's, including the runner-hang retry: a run that produces NO
assertion lines never got far enough to assert, which is the known simulator flake, so retry once
after shutting the simulators down. A real failure prints assertions and exits immediately, so
the retry cannot mask one. Distinct exit code (68 vs 67) so the two gates are tellable apart.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
bjorkert added a commit to Niklasson1/LoopWorkspace that referenced this pull request Nov 8, 2026
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.

2 participants