host: wire Variant C history-image opts through node + worker CLI

Threads compressHistory / historyKeepTail / historyMinPrefix through CliOpts,
env vars (COMPRESS_HISTORY, HISTORY_KEEP_TAIL, HISTORY_MIN_PREFIX), and CLI
flags (--history, --no-history, --history-keep-tail N, --history-min-prefix N).
Same wiring on Workers host. Startup log reports history=on/off.

TrackEvent extended with 4 new fields (collapsed_turns, collapsed_chars,
collapsed_images, history_reason) — emitted only when non-zero/defined.

Defaults: compressHistory=false (opt-in given round-3 MARGINAL rating).
229/229 tests green, 496 KB gzipped.

src/core/tracker.ts

@@ -44,6 +44,18 @@ export interface TrackEvent {
   truncated_tool_results?: number;
   /** Total chars elided by paging across all tool_results this request. */
   omitted_chars?: number;
+  /** Variant C history-image: how many messages got collapsed into the
+   *  prepended synthetic user message this request. Absent when not. */
+  collapsed_turns?: number;
+  /** Variant C: total chars of text serialized into the history image(s)
+   *  before render. Absent when no collapse happened. */
+  collapsed_chars?: number;
+  /** Variant C: number of PNG image blocks emitted for the history. Folded
+   *  into `image_count` too — surfaced separately so dashboards can
+   *  attribute image-count growth to history vs system-slab vs reminders. */
+  collapsed_images?: number;
+  /** Variant C: why the history collapse didn't run (or did). Diagnostic. */
+  history_reason?: string;
   /** Codepoints rendered into images that weren't in the glyph atlas. A
    *  spike here means users are typing glyphs we don't ship — consider
    *  switching ATLAS_PROFILE to `full-bmp`. */
@@ -161,6 +173,18 @@ export function toTrackEvent(ev: ProxyEvent): TrackEvent {
     if (info.omittedChars !== undefined && info.omittedChars > 0) {
       out.omitted_chars = info.omittedChars;
     }
+    if (info.collapsedTurns !== undefined && info.collapsedTurns > 0) {
+      out.collapsed_turns = info.collapsedTurns;
+    }
+    if (info.collapsedChars !== undefined && info.collapsedChars > 0) {
+      out.collapsed_chars = info.collapsedChars;
+    }
+    if (info.collapsedImages !== undefined && info.collapsedImages > 0) {
+      out.collapsed_images = info.collapsedImages;
+    }
+    if (info.historyReason !== undefined) {
+      out.history_reason = info.historyReason;
+    }
     if (info.droppedChars !== undefined && info.droppedChars > 0) {
       out.dropped_chars = info.droppedChars;
     }

src/node.ts

@@ -33,6 +33,11 @@ interface CliOpts {
   compressSchemas: boolean;
   compressReminders: boolean;
   compressToolResults: boolean;
+  /** Variant C history-image compression (opt-in). Marginal savings per
+   *  round-3 spec; only enable once cache topology is verified. */
+  compressHistory: boolean;
+  historyKeepTail: number;
+  historyMinPrefix: number;
   minCompressChars: number;
   minReminderChars: number;
   minToolResultChars: number;
@@ -59,6 +64,13 @@ function parseCli(argv: string[]): CliOpts {
     compressSchemas: envFlag('COMPRESS_SCHEMAS', true),
     compressReminders: envFlag('COMPRESS_REMINDERS', true),
     compressToolResults: envFlag('COMPRESS_TOOL_RESULTS', true),
+    // Variant C history-image: OFF by default. Round-3 spec marks the
+    // savings as MARGINAL (~1% per-call) against HIGH cache-topology risk.
+    // Flip via COMPRESS_HISTORY=1 or --history once telemetry shows the
+    // static-slab cache_control still hits.
+    compressHistory: envFlag('COMPRESS_HISTORY', false),
+    historyKeepTail: Number(process.env.HISTORY_KEEP_TAIL ?? 4),
+    historyMinPrefix: Number(process.env.HISTORY_MIN_PREFIX ?? 10),
     minCompressChars: Number(process.env.MIN_COMPRESS_CHARS ?? 2000),
     // Raised to 10,000 — the per-block break-even point at the current
     // renderer config (Unifont 10px, cell 5×11, 100 cols). The real gate
@@ -85,6 +97,10 @@ function parseCli(argv: string[]): CliOpts {
       case '--no-schemas':     o.compressSchemas = false; break;
       case '--no-reminders':   o.compressReminders = false; break;
       case '--no-tool-results':o.compressToolResults = false; break;
+      case '--history':        o.compressHistory = true; break;
+      case '--no-history':     o.compressHistory = false; break;
+      case '--history-keep-tail': o.historyKeepTail = Number(eat()); break;
+      case '--history-min-prefix': o.historyMinPrefix = Number(eat()); break;
       case '--min-chars':      o.minCompressChars = Number(eat()); break;
       case '--min-reminder-chars': o.minReminderChars = Number(eat()); break;
       case '--min-tool-result-chars': o.minToolResultChars = Number(eat()); break;
@@ -464,6 +480,9 @@ async function main(): Promise<void> {
     compressSchemas: opts.compressSchemas,
     compressReminders: opts.compressReminders,
     compressToolResults: opts.compressToolResults,
+    compressHistory: opts.compressHistory,
+    historyKeepTail: opts.historyKeepTail,
+    historyMinPrefix: opts.historyMinPrefix,
     minCompressChars: opts.minCompressChars,
     minReminderChars: opts.minReminderChars,
     minToolResultChars: opts.minToolResultChars,
@@ -584,7 +603,7 @@ async function main(): Promise<void> {
   server.listen(opts.port, () => {
     console.log(`[pixelpipe] listening on http://127.0.0.1:${opts.port} → ${opts.upstream}`);
     console.log(
-      `[pixelpipe] config: compress=${opts.compress} tools=${opts.compressTools} schemas=${opts.compressSchemas} reminders=${opts.compressReminders} tool_results=${opts.compressToolResults} min=${opts.minCompressChars} placement=${opts.placement} cols=${opts.cols}`,
+      `[pixelpipe] config: compress=${opts.compress} tools=${opts.compressTools} schemas=${opts.compressSchemas} reminders=${opts.compressReminders} tool_results=${opts.compressToolResults} history=${opts.compressHistory} min=${opts.minCompressChars} placement=${opts.placement} cols=${opts.cols}`,
     );
     if (opts.track) console.log(`[pixelpipe] tracking events → ${opts.eventsFile}`);
     else console.log('[pixelpipe] tracking disabled (--no-track or PIXELPIPE_TRACK=0)');

src/worker.ts

@@ -24,6 +24,11 @@ export interface Env {
   COMPRESS_SCHEMAS?: string;
   COMPRESS_REMINDERS?: string;
   COMPRESS_TOOL_RESULTS?: string;
+  /** Variant C history-image compression. OFF by default — round-3 spec
+   *  rated savings as MARGINAL vs HIGH cache-topology risk. */
+  COMPRESS_HISTORY?: string;
+  HISTORY_KEEP_TAIL?: string;
+  HISTORY_MIN_PREFIX?: string;
   MIN_COMPRESS_CHARS?: string;
   MIN_REMINDER_CHARS?: string;
   MIN_TOOL_RESULT_CHARS?: string;
@@ -46,6 +51,12 @@ export default {
       compressSchemas: truthy(env.COMPRESS_SCHEMAS, true),
       compressReminders: truthy(env.COMPRESS_REMINDERS, true),
       compressToolResults: truthy(env.COMPRESS_TOOL_RESULTS, true),
+      // Variant C history-image: OFF by default. Round-3 spec marks the
+      // savings as MARGINAL (~1% per-call) against HIGH cache-topology risk.
+      // Flip via COMPRESS_HISTORY=1 once telemetry confirms safe rollout.
+      compressHistory: truthy(env.COMPRESS_HISTORY, false),
+      historyKeepTail: env.HISTORY_KEEP_TAIL ? Number(env.HISTORY_KEEP_TAIL) : 4,
+      historyMinPrefix: env.HISTORY_MIN_PREFIX ? Number(env.HISTORY_MIN_PREFIX) : 10,
       minCompressChars: env.MIN_COMPRESS_CHARS ? Number(env.MIN_COMPRESS_CHARS) : 2000,
       // Raised to 10,000 — per-block break-even point at current renderer
       // config (Unifont 10px, cell 5×11, 100 cols). Real gate is