Each visible chart row was hardcoded to a fixed pixel height (110px in
advanced mode, 56px-capped in compact) regardless of the widget's actual
tile size - resizing the tile taller left the extra height unused, and
some visible charts required scrolling to reach.
Divide the available height evenly across rows using CSS Grid's
grid-auto-rows: 1fr (the Stack fills the tile via a definite pixel
height, the SimpleGrid grows into it via flex and divides its rows
evenly) rather than computing a per-chart pixel height in JS. Confirmed
via live inspection of a deployed board that Mantine's h prop is not
safe for this: its rem() converter strips a "px" suffix from any string
before checking whether it's a parseable number, so even a literal pixel
*string* passed through h= gets rescaled by rem() * var(--mantine-scale)
exactly like a raw number would - style.height is the only way to set
an exact, un-rescaled pixel value.
A container's outer grid cell and its own inner content grid are sized by
two independent calculations that only agree when the container is
scrollable and expanded. When collapsed, or when non-scrollable content
needs more rows than declared, the inner grid keeps rendering at its old/
content-driven size while the outer cell shrinks or stays fixed, leaving
an invisible-but-space-occupying box that inflates the board's scroll
height into a visible empty gap below the container.
Pass the outer instance's collapsed row count down as a
viewportRowCountOverride so the inner grid matches it exactly when
collapsed, and grow the outer placement's height to the already-computed
content minimum (getContainerMinimumSizes) for non-scrollable containers
so it stays in sync with the inner grid's own growth.
* feat(header): add opt-in auto-hide-on-scroll for the header
New per-user header preference (default off): when enabled, the
header slides away while scrolling down and reappears on scroll-up or
near the top of the page. Disabled while in board edit mode, and
skipped entirely while focus is inside the header. Only applies when
the header is set to visible - the existing hover-to-reveal floating
controls already cover the hidden-header case.
* chore(api): regenerate the checked-in OpenAPI schema
check:openapi was failing in CI - the schema on disk hadn't been
regenerated after the header preferences gained autoHideOnScroll.
Add scripts/generate-openapi-schema.mts (and a generate:openapi
package script) alongside the existing check script so this is a
one-command fix next time the schema drifts.
* docs(users): document the header studio's visibility and auto-hide options
Header customization (drag-to-arrange controls, hide/show, hover-reveal
and pin, auto-hide-on-scroll) had no documentation on the user general
settings page.
* fix(calendar): add cell spacing and scale day content with widget size
* fix(beszel): stop chart panels from leaving blank grid columns and size charts to fit the widget
* fix(widgets): address CodeRabbit review on calendar/beszel scaling PR
- beszel: account for the SimpleGrid row gap and the Stack gap between
the controls row and the grid in the available-height budget, so a
two-row layout no longer overflows the widget.
- calendar: memoize the day-cell scale lookup instead of recomputing
it on every render.
- calendar: don't render the notification indicator row when none of
a day's events have an indicatorColor, which was shifting the date
number up with no visible dot to show for it.
Skipped one finding: the claim that .day's absolute inset:0 covers
monthCell's padding. It doesn't - the containing block for an
absolutely positioned element is the padding edge of its positioned
ancestor per the CSS spec, so the padding correctly insets .day. This
was also visually verified in the browser while building the fix.
* fix(beszel): don't treat availableHeight=0 as omitted in computeChartHeight
A falsy check on availableHeight also caught the legitimate value 0
(widget shorter than its own padding+controls), falling back to the
full 180px height instead of clamping to MIN_CHART_HEIGHT - the exact
overflow this height budget was meant to prevent.
* fix(board): give AppShell root a viewport-height floor when a background is set
The board canvas sizes itself to its content, and AppShell runs in
'static' mode so <main> doesn't force a 100dvh minimum either. That
means a board's background image only covers as much height as the
content needs - collapsing a container (or any board with short
content) leaves flat page background showing below it instead of the
background continuing to the bottom of the viewport, looking like a
big empty void.
Root-caused via a local repro: the grid/reflow system was already
shrinking the collapsed container's own space correctly (verified the
live-measured canvas height matched exactly, not the stale debug
attribute) - the void was purely a background-sizing issue on the
AppShell root, unrelated to the collapse/layout logic itself.
* docs(boards): document background minimum-height behavior
The small refresh-circle overlay that appeared in a widget's top-right
corner during a background refetch wasn't adding useful signal and
just added visual noise. Drop the isWidgetRefreshing detection, the
indicator markup/CSS, and the now-dead query-key/runtime-query
plumbing that only existed to power it.
Updated the lazy-widgets e2e test to drop its assertions on the now
-removed [data-widget-refreshing] marker; the behavior it was really
verifying (cached content stays visible during a background refetch,
which completes cleanly) is still covered by the existing network
-level checks in the same test.
logicalWidth/viewportHeight can go non-positive for a narrow container
on a heavily zoomed-out board, where outerCardInset (which grows as
canvasScale shrinks) can exceed the container's own un-inset size.
That non-positive scale then feeds into a division (Infinity/NaN into
--board-canvas-ui-scale) and is used as zoom directly (invalid or
collapsed content) - flooring it well above zero avoids both.
Removing it in favor of viewportHeight for the vertical fit ratio left
the declaration itself dead code, which the CI lint step flags as an
error rather than a warning.
- Guard canvasScale before dividing by it in outerCardInset, matching
calculateBoardUiScale's own fallback.
- Fix containerContentScale's vertical fit ratio for scrollable
containers: it was derived from the full (all-rows) content height,
which drifts toward 1 as more rows are added regardless of the fixed
inset, under-scaling relative to what the visible viewport actually
needs and clipping the bottom of the visible rows. Use
viewportHeight/viewportRowCount instead, which describe the real
visible card size in both the scrollable and non-scrollable case.
A container's own visible card is inset from its allocated board cell
by the board's standard per-item gap, but items inside it are
positioned in a coordinate space sized to the un-inset column/row
count - so without correcting for that, items (not just the grid's
own box) spilled past the container's right/bottom edge by that
inset amount, worse the busier/more zoomed-out the board is.
Scales the container's inner grid content down to fit the actual card
size, using the same zoom mechanism the board's own top-level canvas
uses for its whole-board zoom, just one level deeper. Corrects
--board-canvas-ui-scale to compensate so icon/text/custom-CSS sizing
inside the container isn't affected - computed as a plain JS number
rather than a CSS calc() referencing the property's own inherited
value, since custom properties don't have a sequential/temporal order
within the same element and doing so is a circular reference that
silently invalidates the whole property for every descendant.
Also fixes the collapsible container's label bar poking above the
card's rounded top border via a negative offset - it now sits flush
inside it.
- Render the chart tooltip via a portal so it isn't clipped by the
card's overflow:hidden (needed to stop the chart line itself
bleeding past its rounded corners).
- Move chartColumns' width breakpoints into a lookup table and
memoize the computed value, instead of an inline ternary.
The chart card and its outer widget shell both had overflow:visible,
letting chart lines bleed past their rounded borders into neighboring
widgets. Compact-mode charts also stretched to fill the widget's full
resized height instead of staying compact like other widgets.
Additionally, the history used to build the chart needs 2+ points to
render a line - it started empty and only got its first real point on
the next poll (up to 30s later), so the chart would flash between a
single-value fallback and the actual graph. Both the initial render
and the first state update now seed with the current value duplicated,
so a flat line renders immediately with no flicker.
ajnart
Thomas
Bergasha