mirror of
https://github.com/Klipper4FlashForge/firmware.git
synced 2026-10-19 03:49:50 -05:00
Oleksandr ShyshatskyiandClaude Opus 5
ba01b8f9f0
`for p in dist/*.tgz` fed the replica the Creator5 package first. That machine's own app_startup.sh globs /mnt/Creator5Pro-*.tgz, found nothing, and fell through to a normal boot -- so all 123 replica tests errored with "could not install the package into a replica", about a package that was built correctly and whose model gate was working exactly as designed. The loop is older than the pytest lane and used to be fine: sim-install.py resolved a machine per package. The replica lane takes PRINTER_IMAGE from test.env, and there is one of those. So the end-to-end runs on the model the replica IS, and a Creator5 end-to-end waits on a Creator5 replica image that nothing publishes yet. Both models are still built and both are still attached to the release; only the install test narrows. upload-artifact moves above that step while I am here, so a run that produced packages hands them over even when the lane fails. The release gate is unchanged: Publish comes after, and a failed step still stops the job before it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01S64uRMdZbj9xPMns2dqMG8