+ - Done in #691. The site's `SearchEntry` is the contract's `SearchEntryView`: any kind, what every kind carries required, the rest optional (FR-003's view, since the search page shows every kind side by side and a template cannot read a field one kind lacks); `IconEntry` is the contract's kind. Two readers changed: the protein forms' key drops a fallback `id` already covers (`id` is the stId, or the dbId, never empty), and an icon's image is found by `id`, the same value as `stId` but never highlighted. From #690's review, the icon library too (#691): its search is highlighted, and asking beta's service for `R-ICO-012449` there answers `stId` as `<span class="highlighting" >R-ICO-012449</span>`, `id` plain, so its image, link, fetch and shown id were built from markup; they take `id` now, as does the pathway browser's icon load. From #691's review (#697): the search page printed a highlighted reference identifier as text, `PTEN (<span class="highlighting" >P60484</span>)` for a search of P60484, seen in the browser against beta's services; it is rendered as its neighbours are (`[innerHTML]`) and reads `PTEN (P60484)`. An icon's alt text and the pathway browser's icon link take `id`, and the interactor page strips a highlighted name as it strips the identifier. Nothing else visible changes; the search, search-error and icon-library end-to-end specs pass (46 and 78 tests with the specs below). The single icon page reads only type changes; no recording opens one (the `gap` T014 closes).
0 commit comments