feedback5 R2: docs and comments follow the split — DESIGN §2 and META list server/match/match/ and server/match/player/ (the classes: constructor + method install), research 09 / 11 name the new homes (Match.watch in match/watch.js, the settle in match/settle.js, the facing changes in player/placement.js / items.js), code comments that pointed at moved code say where it lives now (shared constants, client bandDraft / observe / result / hud / teamPanel / underframe / game-mock, test comments); .git-blame-ignore-revs lists the two pure-move commits. DESIGN §17–§23 history citations, the "before" column and the proposal of research 11, tools/build-data.mjs (Match.js still exports BAND_TURN_SECONDS) and public/js/ui/gameLogic.js (in the client stream R5) are left as they are

test/ui/bandDraft.test.js

@@ -1,8 +1,8 @@
 // Strategy draft 队友已选 (research 09 §5 / §7, DESIGN §14 corrections): a strategy a teammate already picked cannot be
-// chosen again — the server refuses it (server/match/Match.js pickBand → BAD_TARGET '队友已选'), bots re-draw, and the
+// chosen again — the server refuses it (server/match/match/phases.js pickBand → BAD_TARGET '队友已选'), bots re-draw, and the
 // UI marks it (screens/bandDraft.js teammateBands). Automatic assignments (a turn that runs out, a departing seat) give
-// the official default 「华法琳」 only while no teammate holds it, else the first free strategy (Match.js defaultBand; the
-// UI names it: bandDraft.js timeoutBand) — a timed-out turn takes the highlighted band first (g.bandFocus, Match.js
+// the official default 「华法琳」 only while no teammate holds it, else the first free strategy (match/phases.js defaultBand; the
+// UI names it: bandDraft.js timeoutBand) — a timed-out turn takes the highlighted band first (g.bandFocus, match/phases.js
 // timeoutBand; user playtest #4 item 4: one countdown, BAND_TURN_SECONDS per turn, no separate step cap).
 
 import { test, describe } from 'node:test';

test/ui/bosshit-ticker.test.js

@@ -1,5 +1,5 @@
 // Player report after 0.1.0, client part: "隐藏boss还没打就出了造成50%伤害播报". Besides the server's share (fixed in
-// server/match/Match.js, test/match/bosshit-ticker.test.js) the ticker strip plays its queue in order, TICKER_MS (5.2 s)
+// server/match/match/bossRounds.js, test/match/bosshit-ticker.test.js) the ticker strip plays its queue in order, TICKER_MS (5.2 s)
 // per line and up to QUEUE_MAX (4) lines behind: in a headless-Chrome run the Final Assault's "超过20%" / "超过80%"
 // lines played 2–8 s into the Hidden Core with its leader at 96 % / 53 %. A leader-damage line (BOSS_HIT) now plays
 // only during the round it came in, and a player's newer BOSS_HIT line of the round supersedes their older one, queued

test/ui/playtest3.test.js

@@ -2,7 +2,7 @@
 //   2  the own LP drops live while the own battle's enemies enter the blue gate — pendingLoss / liveLp (the settle
 //      rule min(lpCapPerRound, counted leaks), reset when the settled m.private lands, 联防中 during 联防), the team
 //      panel rows (rowLp: own live value, teammates' m.public players[].pendingLp), the LP tower's −N tick, and the
-//      server's pendingLp (server/match/Match.js _pendingLpView)
+//      server's pendingLp (server/match/match/views.js _pendingLpView)
 //   3  the temp overflow row (临时整备区): the ready button's visible reason, the row's frame geometry (tempRowFrame)
 //   8  the operator's own effect (特质 / garrison) right under the detail card's header (CHESS_SECTIONS, rendered order)
 //   9  no spinning busy indicators next to texts: no wait / progress cursors, the button's busy bar instead of a