Recover blocked searches, fall through to the next peer, rebuild the UI

Measured against a live instance: 47 failed jobs against 33 imported. Three
independent causes, all of which ended the request.

The Soulseek server filters some artist names. "Michael Jackson Billie Jean"
and bare "Michael Jackson" both return zero responses; "Billie Jean" returns
1083 files. The Weeknd, Kanye West and Frank Ocean behave the same way. Since
pick_and_enqueue sent exactly one query of `artist title`, every request for a
filtered artist failed reporting "searched 0 files" -- 28 of 29 occurrences.
search_queries() now walks a ladder, loosest last, ending at the bare title,
and escalates only when a search underdelivers so the common case still costs
one search. Dropping the artist from the query does not mean ignoring it:
match_ratio() looks for it in the folder path, where an organised share puts it.

A rejected transfer ended the request. pick_and_enqueue kept a running maximum
and discarded every runner-up, so "Transfer rejected: File not shared." -- a
per-file condition where candidate two almost always works -- was terminal, as
were timeouts and slskd's 500 on an unreachable peer. The ranked list is now
persisted on the job row, ordered to put a different peer next, and
try_next_candidate() advances through it. A job fails only once the list is
exhausted, and then reports every peer it tried and what each one said.

norm() destroyed every non-ASCII character, so "Beyonce" never matched
Beyonce, nor Bjork Bjork, nor Dont Don't. It now folds accents and drops
apostrophes while leaving & as a separator.

Also: availability outranks format in score(), because a FLAC behind 5000
queued users used to beat an immediately available MP3; transfers that stall or
that slskd forgets are abandoned instead of sitting in `downloading` forever;
searches are deleted only once finished, since deleting one mid-flight makes
slskd throw an optimistic-concurrency error and lose the record; a failed
search POST is retried rather than failing the job on one blip.

The UI moves to a sidebar shell -- rail at tablet, bottom bar on phones -- with
neutral chrome and the sampled artwork colour confined to hero areas rather
than repainting the whole interface. Adds URL routing, a primitive set under
components/ui, and a request sheet showing the attempt ladder. Fixes
hover:bg-glass-hi, which was never a valid class, so track rows and the search
field had no hover or focus state at all; a reduced-motion path that could
strand elements at opacity 0; the login page's hardcoded palette; and the
missing 404 and error boundaries.

tests/test_acquisition.py covers this path against a faked slskd. It had none.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

web/lib/motion.ts

@@ -0,0 +1,84 @@
+"use client";
+
+import { useReducedMotion, type Transition, type Variants } from "framer-motion";
+
+/**
+ * One motion vocabulary.
+ *
+ * There were four different stagger formulas and five entrance offsets across
+ * the components, so nothing quite moved like anything else. More importantly,
+ * the `prefers-reduced-motion` block in globals.css only governs CSS: Framer
+ * animates via the CSSOM from JS, so every entrance, the aurora blobs and the
+ * pulsing status dots all ran at full strength for people who had explicitly
+ * asked them not to. `useMotionOK()` is the fix, and every animated component
+ * consults it.
+ */
+export function useMotionOK() {
+  return !useReducedMotion();
+}
+
+/** Apple's motion reads as weight, not duration: overshoot slightly, settle. */
+export const spring: Transition = { type: "spring", stiffness: 420, damping: 34 };
+export const springSoft: Transition = { type: "spring", stiffness: 260, damping: 30 };
+
+export const rise: Variants = {
+  hidden: { opacity: 0, y: 8 },
+  shown: { opacity: 1, y: 0 },
+};
+
+export const fade: Variants = {
+  hidden: { opacity: 0 },
+  shown: { opacity: 1 },
+};
+
+/**
+ * Capped, so item 400 of a long list does not wait eight seconds for its turn.
+ */
+export function stagger(i: number, step = 0.025, max = 0.3) {
+  return Math.min(i * step, max);
+}
+
+/**
+ * Entrance props for a list item, honouring the reduced-motion setting.
+ *
+ * Note what the reduced-motion branch does NOT do: return `{}`. Dropping the
+ * props entirely removes `animate` while `initial` has already painted
+ * opacity 0, and the element stays invisible forever. `useReducedMotion()`
+ * also reports false on the first render and flips after mount, so that lands
+ * exactly on the people who asked for less motion and shows them a blank page.
+ * Instead the element is mounted already in its final state.
+ */
+export function item(i: number, ok: boolean) {
+  if (!ok) return { initial: false as const, animate: "shown" as const, variants: rise };
+  return {
+    initial: "hidden" as const,
+    animate: "shown" as const,
+    variants: rise,
+    transition: { ...springSoft, delay: stagger(i) },
+  };
+}
+
+/**
+ * Entrance for a whole view. Deliberately translate-only, with no fade.
+ *
+ * This wrapper contains the entire screen, so anything that can leave it at
+ * opacity 0 leaves the app looking broken rather than looking unanimated. A
+ * background tab throttles rAF and freezes an in-flight animation part way —
+ * it resumes on focus, but a stalled fade is a blank page and a stalled slide
+ * is a few pixels of offset nobody notices. Fades stay on individual items,
+ * where a stall is cosmetic.
+ */
+const slide: Variants = {
+  hidden: { y: 8 },
+  shown: { y: 0 },
+};
+
+export function panel(ok: boolean) {
+  if (!ok) return { initial: false as const, animate: "shown" as const, variants: slide };
+  return {
+    initial: "hidden" as const,
+    animate: "shown" as const,
+    variants: slide,
+    transition: { duration: 0.22, ease: [0.22, 0.9, 0.28, 1] as const },
+  };
+}