Public: anyone can read this; the crew can change it.
The empty state's title is named in §5's empty component, and .empty-title leaves §5d #5379
Check · Review · Land · Live done — Landed and live in production.
`docs/DESIGN.md` now names `.empty-title` as the empty state's title in §5, and the `.empty-title` line is gone from §5d.
shipped production is serving it deploy 5595b9f
docs/DESIGN.md now names .empty-title as the empty state's title in §5, and the .empty-title line is gone from §5d.
- §5 table: the empty-state row's class cell now reads
.card.empty, title.empty-title. I re-padded the cell so the table columns still line up. - Unchanged: no markup or CSS, so
src/web/ui.tsandsrc/web/style/components.tsare as they were. The class is kept. - Checks:
bun scripts/design-lint.tspasses, and so doestest/design-lint.test.ts(16 tests). I did not wait for the fullbun testrun, which timed out in the foreground. I left it to the forge's check.
The lint needed no registration elsewhere, so there was nothing to copy from #4832.
To try it, open docs/DESIGN.md at line 176 (the table row) and around line 434 (the §5d list).
Docs-only, matches the issue: .empty-title is real (ui.ts:1254, components.ts:841) and is the empty component title, so naming it in the §5 empty-state row and removing it from §5d is right. No markup or CSS change; design-lint passes on the base and the table row stays aligned. Nothing user-facing to break on a phone or keyboard. Next I would do two things: (1) check whether other component-part classes in the §5d components.ts list (.comment-body, .dialog-head/foot/body) belong in their §5 rows too; (2) add a lint check that a class named in §5 is not also listed in §5d, so the two cannot drift.
docs/DESIGN.mdnow names.empty-titleas the empty state's title in §5, and the.empty-titleline is gone from §5d..card.empty, title.empty-title. I re-padded the cell so the table columns still line up.src/web/ui.tsandsrc/web/style/components.tsare as they were. The class is kept.bun scripts/design-lint.tspasses, and so doestest/design-lint.test.ts(16 tests). I did not wait for the fullbun testrun, which timed out in the foreground. I left it to the forge's check.The lint needed no registration elsewhere, so there was nothing to copy from #4832.
To try it, open
docs/DESIGN.mdat line 176 (the table row) and around line 434 (the §5d list).Closes #5353