# Browser evidence

These are transcribed observations from 12 actual local browser-tool calls against `output/site/`. Each call starts a fresh browser; actions within a row were performed together. The tool uses a trusted server implementing the original `/api/leads` behavior. All 12 returned `overflow: false` and `errors: []`. Observations are text/control reports, not screenshots.

Source requirements: `inputs/brief.md`, **Workshop site acceptance brief**. Source content/identity: `site/index.html` (title, header, workshop copy and form/aside); `site/style.css` (dark-blue/orange colors).

## Calculator and valid submissions

| ID | Width | Actions | Observed result |
|---|---:|---|---|
| B01 | 1280 | Load page with default 1 seat | `CAD 268.80`; name/email empty and invalid, seats valid |
| B02 | 390 | Fill Mobile Tester, mobile@example.test, 3 seats; click Request a place | `CAD 725.76`; all inputs valid; `Thanks! Your workshop request is saved.` |
| B03 | 1280 | Fill Desktop Tester, desktop@example.test, 2 seats; click Request a place | `CAD 537.60`; all inputs valid; same success message |
| B04 | 390 | Fill seats with 8 | `CAD 1935.36`; seats valid |

All four examples match `inputs/brief.md`, **Workshop site acceptance brief**:

| Seats | Subtotal | Discount | After discount | Tax (12%) | Displayed CAD total |
|---:|---:|---:|---:|---:|---:|
| 1 | 240.00 | 0.00 | 240.00 | 28.80 | 268.80 |
| 2 | 480.00 | 0.00 | 480.00 | 57.60 | 537.60 |
| 3 | 720.00 | 72.00 | 648.00 | 77.76 | 725.76 |
| 8 | 1920.00 | 192.00 | 1728.00 | 207.36 | 1935.36 |

## Invalid submissions

Every row below included a click on Request a place and showed:
`Please check your name, email and seats (a whole number from 1 to 8).`

| ID | Width | Inputs | Observed validation/calculator |
|---|---:|---|---|
| B05 | 390 | Invalid Seats / invalid@example.test / blank seats | Seats invalid; `Enter 1–8 seats` |
| B06 | 390 | Same name/email / 0 seats | Seats invalid; `Enter 1–8 seats` |
| B07 | 390 | Same name/email / -1 seats | Seats invalid; `Enter 1–8 seats` |
| B08 | 1280 | Same name/email / 1.5 seats | Seats invalid; `Enter 1–8 seats` |
| B09 | 1280 | Same name/email / 9 seats | Seats invalid; `Enter 1–8 seats` |
| B10 | 390 | Three spaces for name / name@example.test / 1 seat | Name invalid; `CAD 268.80` |
| B11 | 1280 | Email Tester / name@example / 1 seat | Email invalid; `CAD 268.80` |
| B12 | 390 | Empty name and email / default 1 seat | Name and email invalid; `CAD 268.80` |

## Storage evidence

After all browser calls, read `output/site/submissions.jsonl`: it contained exactly the two B02/B03 records and no others. The records were copied verbatim to `output/browser-records.jsonl` (lines 1–2). This verifies the two successful browser signups were stored and the invalid browser attempts added no records in aggregate. No per-request network log was available.

Test records were then removed from `output/site/`, so the shipped site starts without sample leads; the server creates `submissions.jsonl` on the first successful signup (`site/server.py`, `Handler.do_POST`).

## Scope

Desktop/mobile text reports exposed labels and controls and reported no horizontal overflow. CSS review confirms a single-column layout below 760px, 18px base text and full-width padded inputs/buttons. No screenshot, physical-device, screen-reader, keyboard-only, offline/network-failure or cross-browser-engine check was performed. No browser assertions were made about those untested behaviors.
