ci: widen drift-trio convergence windows for starved CI boxes

drift-trio-scenarios S4/S5 flaked 4× across this week's CI runs (three wasm
instances + full-suite load on the runner) while passing 77/78 locally under
stress — and the one local miss carried zero guard beacons, i.e. the same
under-load convergence shortfall, not a code path. Same treatment as the
follow spec: condition-based polls keep their shape, windows grow to what a
starved box actually needs (inline S4/S5 polls 20s→60s, waitAllContain
30s→90s, settleConverged 30s→90s). Local convergence stays ~1s.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019SE4o46Lnq3hF574FFq8x4

tests/kicad/utils/trio.ts

@@ -419,7 +419,7 @@ function firstDiff(a: string, b: string): string {
  * Poll until every tab's silent save is byte-identical. The convergence gate
  * between scenario steps — bounded poll, no blind sleeps (tests/TESTING.md).
  */
-export async function settleConverged(trio: TabSet, cfg: ToolCfg, timeout = 30000): Promise<void> {
+export async function settleConverged(trio: TabSet, cfg: ToolCfg, timeout = 90000): Promise<void> {
   await expect
     .poll(
       async () => {