rinkhals-community/Rinkhals 20260726_01_test

Rinkhals 20260726_01_test

on GitHub

This build exists mainly so people who reported issues can retest. If you are one of them, your issue is listed below with exactly what to check.

If you reported an issue, please test this

• Issue #85 - bed mesh calibration hangs at the first probe point (KS1). Fixed, and reproduced on hardware first. The calibration script never homed the printer; GoKlipper then blocks silently inside BED_MESH_CALIBRATE instead of reporting "must home first", and because gcode is forwarded with no timeout nothing ever surfaces an error. Test: Mainsail > Heightmap > Calibrate, and confirm it probes the whole mesh and finishes instead of parking at X5 Y5.
• Issue #86 - camera stopped working after updating to 20260716_02 (KS1). Possible cause addressed. In 20260716_02 the web portal's app enable/disable toggle only set a flag and never actually started or stopped the app, so toggling the camera app off broke it at the next reboot and toggling it back on did not recover it. That is fixed here. Test: check that mjpg-streamer is enabled, then confirm the camera stream works. If it was already broken, this build plus a reboot should recover it.
• Issues #87 and #59 - print start kills GoKlipper. A related defect is fixed, though we have not proven it is your root cause. Rinkhals strips a ; filament_colour_type line that newer OrcaSlicer emits and that GoKlipper mis-parses, but that strip only ever ran for multi-tool files. Single-colour prints were handed to GoKlipper untouched, so the protection could never run for them. It now applies to any file containing the line. Test: the exact print that failed before. If it still crashes, please attach a fresh debug bundle - other/gklib.log right after the crash is the file that matters most.
• Issue #89 - Spoolman / MMU_GATE_MAP. The requested feature is not here yet. What is fixed is a shipping bug found while investigating it: Rinkhals' own wrapper around set_active_spool raised a TypeError whenever Moonraker tried to deactivate a spool, which affects any install with a real [spoolman] configured.

Other changes since 20260716_02

• Disabling an app in the web portal now actually stops it, and enabling one starts it, instead of only flipping a flag that took effect at the next reboot.
• System Logs now shows the live Rinkhals log. The Rinkhals tab was reading an early-boot file that stops being written once startup finishes, so it always looked frozen.
• App Store now shows the version it will actually install, and the touchscreen App Store can update an installed app instead of only removing it.
• Debug bundles now report the correct remote-control mode. The log line printed the previous value, so bundles could say cloud while the printer was in lan mode - this sent issue triage down the wrong path more than once.
• mjpg-streamer no longer risks running two monitor processes at once. The start guard had a race, and because each monitor restarts the streamer by killing it by name, duplicates would kill each other's streamer and the camera would flap.

Notes

• The bed mesh fix adds a homing step at the start of calibration, so calibration now takes slightly longer and homes even if you just homed.
• GoKlipper's BED_MESH_CALIBRATE turns out to run its own preheat and wipe internally, so part of our preamble is redundant. Left in place for now to keep the change small.

Report findings to https://github.com/rinkhals-community/Rinkhals/issues with a debug bundle.

Full Changelog: 20260716_02...20260726_01_test