timescale.info#214 · Remove the redundant Exact rational bounds and offline Licenses sections (#213)
- timeline-app
- pull request
Closes #213. ## Exact rational bounds The **Exact rational bounds** section under the timeline is gone. Its own left and right fields and **Go to exact bounds** repeated what the main bounds already do: - **Time display → Exact rational** makes **Left bound** and **Right bound** show and accept exact fractions; - in any display, those fields keep the exact coordinate when left unchanged. The README, the guide and the time-presentation doc now say to use **Exact rational** instead. Tests and benchmarks read the exact view from the timeline stage, which now carries it as `data-view-left` and `data-view-right`. Those that navigated through the old fields (`test/desktop-bulk.mjs`, `test/label-motion-browser.mjs`, `bench/ui.mjs`, `bench/capacity.mjs`) now type into the main bounds and press **Go**, as a person would. The label-motion test's Gregorian timeline types timestamps such as `2026-10-04T23:59:59Z`. ## Licenses (offline file) The offline file's separate **Licenses** section is gone. It repeated the GPL, which the legal footer already has. Its other content, the full license texts of the libraries bundled into the editor (rational-ordered-map, fraction.js), now follows the footer's **Third-party notices** (`scripts/legal.mjs`), since those texts must stay with every copy. That also adds them to the website's `legal.html`. ## Tests Run locally: - `npm test` (299), typecheck and Prettier. - `test/browser.mjs`: passes in Chromium and Firefox. - `test/offline-browser.mjs`: passes in all three browsers. - The platform test with its browser part. - `test/desktop-bulk.mjs`, against a freshly built desktop app. In my Docker runs, `test/browser.mjs` fails in WebKit in the drafts check. A query cut off by the test's navigation is reported as "access control checks". Unchanged `main` fails the same way there, and CI's WebKit job passed on main's last PR, so it looks like a local timing difference. CI will show whether it reproduces.