mirror of
https://github.com/Kpa-clawbot/meshcore-analyzer.git
synced 2026-07-20 21:40:56 +00:00
626900a22a
Red commit: 91fc49f98a (CI run: pending
until pushed)
**Partial fix for #1668 (M3 of 6).**
M2 cleared ~85% of BLOCKER contrast violations (v3.9.1). M3 addresses
the
M1-audit `thin-small` findings: chips, badges, table cells, and meta
labels
where `font-size < 14px AND font-weight < 500` made text hard to read
for
the operator regardless of contrast — the original "the typography
sucks"
complaint that drove this issue.
## Design (operator-locked)
- Body text floor: **14px**
- Chip / badge / meta floor: **12px AND weight ≥ 500** (or 14px if 400)
- Visual hierarchy preserved — H1/H2/H3 untouched
- No palette changes (M2 owns colors) · no layout (M4 owns that)
## What changed
New `:root` weight tokens: `--fw-{normal,medium,semibold,bold}`.
| Selector | Before (px/weight) | After |
|---|---|---|
| `.nav-link` | 12.8 / 400 | 12-14 fluid / **500** |
| `.tab-btn` | 13 / 400 | 13 / **500** |
| `.alab-pkt` (audio-lab) | 12 / 400 | 12 / **500** |
| `.ch-item-time` | 11 / 400 | **12** / **500** |
| `.ch-item-preview` | 12 / 400 | **13** / **500** |
| `.payload-bar-label` | 12 / 400 | 12 / **500** |
| `.stat-label` | 12 / 400 | 12 / **500** |
| `.col-hidden-pill` | **10** / 700 | **12** / 700 |
| `.skew-badge` | **10** / 600 | **12** / 600 |
| `.filter-group .btn` | 12 / 400 | 12 / **500** |
| `.timestamp-text` (new explicit rule) | inherits 12 / 400 | **13** /
**500** |
| `.data-table` (incl. mobile override) | 12-11 / 400 | **13** / **500**
(mobile 12) |
| `.mono` | 12 / 400 | **13** / **500** |
Estimated thin-small violations cleared: **~5,800 of 6,313 MAJOR** in
the
M1 dataset (timestamps + table cells + nav links are the bulk).
## Letsmesh bar
Letsmesh ships 12.5-15px paired with 600 weight on chips — our 12px+500
floor is one notch lighter but matches the operator-locked spec.
## Tests
TDD: `test-issue-1668-m3-typography.js` (new) parses `style.css` +
`audio-lab.js`, computes effective font-size/weight per selector with
CSS cascade resolution, and asserts the floor on 12 high-impact
selectors.
Red commit fails 10/12 on master; green commit makes all 12 pass. Anti-
tautology verified: reverted `.skew-badge` bump → test fails on the
assertion (not a build error) → restored → test passes.
## Verified on staging (hot-patch)
Computed styles AFTER patch: `.timestamp-text` 13/500, `.skew-badge`
12/600, `.nav-link` 12.8/500.
## Next
- M4 — per-route polish + map legend
- M5 — CI gate that re-runs the M1 probe and fails on regressions
- M6 — A/B verification with the operator
---------
Co-authored-by: openclaw-bot <bot@openclaw.local>