|
|
@@ -36,6 +36,7 @@ All notable changes to Bambuddy will be documented in this file.
|
|
|
- **Schedule Calendar Picker Opens Off-Screen** — Clicking the calendar icon in the print modal's scheduled mode opened the native date picker at the bottom of the viewport instead of near the date field. The hidden `datetime-local` input used `sr-only` positioning which anchored the picker off-screen. Fixed by positioning the hidden input inside the date field's container.
|
|
|
- **SpoolBuddy Kiosk Display Blanking and Crashes** — The kiosk Chromium flags added in 0.2.2.2 caused display instability: `--js-flags=--max-old-space-size=128` crashed the V8 renderer when heap exceeded 128 MB, `--enable-low-end-device-mode` aggressively killed GPU rendering surfaces, and resetting `CHROMIUM_FLAGS` discarded the Pi's GPU defaults (`--enable-gpu-rasterization`, ANGLE/GLES) creating an unstable mixed CPU/GPU rendering path. Fixed by removing both flags, appending kiosk flags to Pi defaults instead of replacing them, adding a `wlr-randr` keep-alive loop to prevent display blanking, and adding `<screenBlankTimeout>0</screenBlankTimeout>` to the labwc config.
|
|
|
- **Sidebar Bottom Icons Cut Off With Smart Plugs** ([#862](https://github.com/maziggy/bambuddy/issues/862)) — Adding smart plug buttons to the sidebar caused the bottom icon row to overflow and get partially cut off. The footer section could be compressed by the flexbox layout when the navigation area grew. Fixed by preventing the footer from shrinking, allowing the expanded icon row to wrap, and adding scroll overflow to the collapsed sidebar icon stack.
|
|
|
+- **AMS History Cleanup Crash Every ~24 Hours** — The periodic cleanup of old AMS sensor history entries failed with "can't compare offset-naive and offset-aware datetimes". The cleanup cutoff used `datetime.now(timezone.utc)` (timezone-aware) but the `recorded_at` column stores naive datetimes via SQLite's `func.now()`. The mismatch caused a TypeError when SQLAlchemy processed the comparison. Fixed by using a naive UTC datetime for the cutoff. The error only appeared once per ~24h because the cleanup runs every 288 recording cycles (288 × 5 min = 24h).
|
|
|
|
|
|
## [0.2.2.2] - 2026-03-27
|
|
|
|