You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
fix(ui): pin the box height of single-line chips so leading cannot move them #1879
Astryx sizes a chip's box and its line box off the same 4px scale, so the two agree by construction: Badge and Kbd set height: var(--spacing-5) (20px) with display: inline-flex; align-items: center, and carry --text-supporting-leading, which is also 20px. The leading cannot move the box because the box is pinned to the value the leading already has.
Several Maka chips pin neither. They let the line box decide the height, so any leading change is a layout change. Measured live during the leading convergence (#1878) by applying the new leading in the running document and remeasuring the box:
selector
box before
box after
parent
.maka-message-time-inline
12px
20px
unchanged
.maka-nav-kbd
18px
20px
+2px
.maka-plan-card-schedule
18px
20px
+2px
.maka-plan-card-run
20px
20px
unchanged — already pinned
.maka-plan-card-run is the shape the others should have: its height is fixed, so the same leading change moved nothing.
Scope
Give the single-line chip family the Astryx shape — a pinned height off the spacing scale plus inline-flex centring — so leading and box height stop being two independent decisions. Covers the former line-height: 1 sites (kbd, tags, button labels, counts, timestamps) and the 12px badges.
Coverage limit
The table is what six e2e fixture scenarios render. Chips on surfaces no fixture boots (.plan-proposal-step-number, .plan-execution-step-marker, .maka-composer-skill-trigger-count, .maka-composer-revision-notice, .settingsUsageRecordCount, .settingsUsageDetailToggle, .maka-sidebar-update-button) were not measured live and need the same check before being changed.
Context
Not folded into #1878, which was scoped to leading only: where a box grew there, it grew, and the PR says so rather than absorbing it into padding.
One correction to the issue's premise, from measuring before changing: .maka-plan-card-run declares no height — its 20px comes from a sibling Badge, not a pin — and it must stay unpinned, since a long run message takes it to 220px at the 480px window floor. Several of the "Coverage limit" chips are also already pinned, just to off-ruler literals or to min-height, which leading can still push past. Full measurements in the PR.
Activity
Astro-Han commentedon Aug 2, 2026
Working on this in
fix/chips-pin-box-height.One correction to the issue's premise, from measuring before changing:
.maka-plan-card-rundeclares noheight— its 20px comes from a sibling Badge, not a pin — and it must stay unpinned, since a long run message takes it to 220px at the 480px window floor. Several of the "Coverage limit" chips are also already pinned, just to off-ruler literals or tomin-height, which leading can still push past. Full measurements in the PR.fix(ui): pin the box height of single-line chips to the control ruler
fix(ui): move the single-line chip family onto Astryx Badge
fix(ui): derive the chip population from shape alone and assert every…
test(desktop): close the guard holes three re-reviews found in the ch…
fix(ui): move the single-line chip family onto Astryx Badge (#1881)
refactor(desktop): componentize the settings page container + residua…
refactor(ui): round the update button and drop its glyph's second ring
refactor(ui): give the update action a slot instead of a row of its o…
chore(desktop): drop tokens nothing reads