cesnimda
18 lines
3.9 KiB
Markdown
18 lines
3.9 KiB
Markdown
# P9-02 visual and responsive evidence
|
||
|
||
The33-route matrix generates396 actual native screenshots:320/768/1440px × English/Bokmål × light/dark. [Route rows](route-results.json) retain applied document language/theme, root width, headings and native read status. [Screenshot index](screenshots.json) identifies full-page originals with SHA256; [contact-sheet index](evidence/index.json) identifies33 overview sheets. Each sheet contains12 first900px crops; it does not show every full-page state.
|
||
|
||
A separate read-only AI reviewer actually inspected all33 original route sheets (396 thumbnail crops). Full-size followups inspected Admin768NB/light, CV320NB/dark and job-detail320EN. The supposed job-detail clipping was a thumbnail artefact; deliberate horizontal tab/progression scrollers remain. No gross blank/overlap/unusable first-viewport layout was otherwise confirmed in those sheets. This is agent image inspection, not human acceptance or exhaustive glyph/contrast/readability review.
|
||
|
||
The reviewer confirmed overlapping public-page h2 text in full EN/light and NB/dark320px originals. P9-V01 now sets lineHeight1.4 on the existing four owners. An unchanged native text-rectangle assertion fails on the original and the insufficient1.2 attempt, then passes both language cases over16 width/theme vectors at320/390/768/1440. Independent inspection of final320EN/light and NB/dark full-page screenshot crops confirms the heading separation.
|
||
|
||
The initial Reminders and Profile matrix captures were loading-only and cannot qualify their body layouts. The matrix now waits for Reminders' settled region and the seeded Profile username. The reviewer inspected the eight corrected320px EN/NB light/dark originals: actual empty-state/forms were visible, with no clipping/overlap identified. Additional final-width inspection must be recorded when actually performed; do not infer it from screenshot generation alone.
|
||
|
||
The prior isolated review inspected all 33 regenerated sheets (396 first-900px crops) and nine full-page originals: `/`, `/reminders` and `/profile`, each at 320 EN/light, 768 NB/light and 1440 NB/dark. Overview inspection missed a separate bottom CTA-title overlap; original-detail review then confirmed P9-V02 at `LandingPage.tsx:261`. All 12 old homepage originals were checked at original detail; no other ink collision was confirmed in hero/feature/step/plan/CTA body text. The CTA now has proportional line height; extended native geometry fails both languages before and passes both after over 16 vectors. Earlier results remain retained as superseded, not silently rewritten.
|
||
|
||
Final `phase9-cta` review actually inspected all 33 latest contact sheets, the same nine latest full-page originals, and native-width bottom-650px CTA crops from all 12 latest homepage originals. The index points to the final `phase9-cta-results` paths. CTA title/body/button remain visibly separated and contained across all 12; Reminders shows settled empty content and Profile reaches its sessions section. No other gross overlap/clipping/blank/loading-only defect was observed in this exact scope. Tall full-page images were scaled layout overviews; this is not blanket readability, human/device/assistive-technology or production acceptance.
|
||
|
||
Existing ordinary browser fixtures cover populated/empty/loading/error/long-translation states, dialog recovery, dense Jobs/Board, long Career/CV, saved-settings variants and the Dashboard editor. Their exact scope is limited to the executable scenarios; this396-route cross-product is not a cross-product of every state. Intentional grid/document scrollers preserve content; fixed CV paper geometry was not changed to fit the screen.
|
||
|
||
Real software keyboards, physical touch/drag, actual zoom/text enlargement, alternate browsers, screen readers and professional Bokmål review remain in the [human checklist](../p9-03/MANUAL_CHECKLIST.md). Automated focus/axe/overflow assertions and emulated touch are recorded separately.
|