Audit 6.2 (t4:292-293): active filter pills are brand-teal tinted with
a 15px close glyph and radius 10 -- was Material's own Chip theming.
Audit 6.3 (t4:294-295): an "Idioma" entry chip is now always reachable
once a search is active -- there was no standalone entry point for
language filtering before (only bundled inside "Filtros"). Opens the
same existing filter sheet rather than a new idioma-only picker.
Audit 6.4 (t4:299): the results-count line now uses eyebrowLabel
styling -- was labelLarge (14/w800).
Audit 13.3/13.4 (t4:649-651): the search-results loading skeleton
gains the missing "BUSCANDO EMISORAS..." eyebrow, and its row gap is
now 4px, not 10.
Audit 13.5/13.6 (t4:657-668): a new purpose-built _TarjetaSinResultados
replaces PluriEmptyState ONLY in the search-no-results branch (that
shared widget is left untouched for its other unrelated call sites --
favorites empty state, the discovery grid). The title now quotes the
typed query ("Sin resultados para <<jazzz>>"), and the clear-filters
pill sits INSIDE the card. New ARB key searchNoResultsForQueryTitle.
Item 6.1 (compact 60px active-search header replacing the persistent
glass pill) is NOT implemented: existing tests confirmed the quality
(bitrate) filter has NO entry point anywhere in the app besides the
header's "Filtros" pill (added first via _abrirYSeleccionarCalidad
while a country filter is ALREADY active, i.e. mid active-search).
Hiding PluriRootHeader during active search would make quality
filtering unreachable -- a real capability regression the task's own
"presentation changes, capability does not" principle forbids.
Re-plumbing where that filter lives is a bigger change than this
visual-fidelity pass should make unilaterally.
FreeTLab
42d35a2541