Version 1.0.2 — English passage lookup

English text now uses the same Read text and Word list views as Latin. Select a word in the displayed passage to see its dictionary entries. Original spelling, punctuation, line breaks and repeated occurrences remain visible.

• Each English word is queried separately. A first word with no match, such as the in the queen, no longer prevents the remaining words from being queried.
• No-match results and original WORDS errors are shown for their affected word; other words retain their results. Keyboard selection and view switching work in both languages, including when saved offline.
• English prompts now accept a word, phrase or sentence. The existing 2,000-word and 20,000-character input limits remain in place.

Application 1.0.2 retains frozen engine 1.0.0, all 28 locked backend paths, dictionary data, native per-word ranking, six-entry trimming, API and CLI behavior. English singular/plural normalization is not included. The browser guide describes spelling support and the browser-only english-passage-v1 structured wrapper.

Verification includes 62 portable tests, the frozen backend guard, Chromium and WebKit passage/list interaction, unknown/error isolation, input boundaries, all 8 raw browser validation cases, saved offline cold restart and upgrade from 1.0.1. The complete source reproduces the runtime and source archive byte-for-byte. Automated checks do not claim physical-device installation testing.

This application patch is distributed through GitHub and the website. No new Zenodo archive is requested; the 1.0.0 archive remains the preserved engine milestone. Citation metadata identifies this patch's artifacts.