billion-context-pi 0.1.59 → 0.1.60-pr.337.131

package/dist/config.d.ts

@@ -206,6 +206,12 @@ export interface AdapterConfig {
      *  disables) or a ThrottleRetryConfig object. Default: enabled, 10 retries,
      *  60s exponential base capped at 300s per kick. */
     throttleRetry?: boolean | ThrottleRetryConfig;
+    /** Cap on the output-headroom reservation as a fraction of the context
+     *  window: reserved = min(model.maxTokens, pct * window). Accepts a ratio
+     *  (0.25) or percent string ("25%"). Default: 0.25. Set 0 to disable the
+     *  reservation entirely; >= 1 restores the legacy full-capability
+     *  reservation (issue #207). */
+    outputHeadroomMaxPct?: number | string;
     /** Generic tool-call repetition guard (see RepetitionGuardConfig). Accepts a
      *  boolean shorthand (`false` disables) or an object. Default: enabled,
      *  warn=3, abort=5. Stops greedy small models looping on byte-identical

package/dist/index.js

@@ -4192,10 +4192,10 @@ var SessionStateStore = class {
   }
   async save(state, sessionFile, sessionId) {
     const file = stateFileFor(sessionFile);
-    if (!file) return;
     const key = cacheKey(sessionFile, sessionId);
     const liveRefOrigins = this.cache.get(key)?.liveRefOrigins ?? [];
     this.cache.set(key, { state, liveRefOrigins });
+    if (!file) return;
     const dir = path2.dirname(file);
     await fs.mkdir(dir, { recursive: true }).catch((e) => {
       logError("state", { event: "save-mkdir-failed", dir, error: e instanceof Error ? e.message : String(e) });
@@ -4462,6 +4462,7 @@ var KNOWN = /* @__PURE__ */ new Set([
   "compress",
   "displayUsage",
   "throttleRetry",
+  "outputHeadroomMaxPct",
   "prompts",
   "acknowledgePromptsRisk"
 ]);
@@ -4699,19 +4700,25 @@ function toTokenNumber(raw) {
   const n = Number(raw.replace(/,/g, ""));
   return Number.isFinite(n) && n >= 1e3 ? n : void 0;
 }
-function reserveOutputHeadroom(window, maxOutput) {
+var DEFAULT_OUTPUT_HEADROOM_MAX_PCT = 0.25;
+function resolveOutputHeadroomCap(value) {
+  return value === void 0 ? DEFAULT_OUTPUT_HEADROOM_MAX_PCT : parsePercent(value);
+}
+function reserveOutputHeadroom(window, maxOutput, capPct = 1) {
   if (Number.isFinite(window) && window > 0 && Number.isFinite(maxOutput) && maxOutput > 0 && maxOutput < window) {
-    return window - maxOutput;
+    const cap = Number.isFinite(capPct) ? Math.max(0, Math.min(capPct, 1)) : 1;
+    const reserved = Math.min(maxOutput, cap * window);
+    return reserved > 0 ? window - reserved : window;
   }
   return window;
 }
 function shouldReserveOutputHeadroom(api) {
   return api !== "anthropic-messages";
 }
-function applyOutputHeadroom(config, model) {
+function applyOutputHeadroom(config, model, capPct = 1) {
   const maxOutput = model?.maxTokens ?? 0;
   if (shouldReserveOutputHeadroom(model?.api)) {
-    const reserved = reserveOutputHeadroom(config.modelContextLimit, maxOutput);
+    const reserved = reserveOutputHeadroom(config.modelContextLimit, maxOutput, capPct);
     if (reserved !== config.modelContextLimit) return { ...config, modelContextLimit: reserved };
   }
   return config;
@@ -9851,6 +9858,9 @@ async function handleCompress(args, runtime, ctx, toolCallId) {
     logWarn("compress", { sid: ctx.sessionManager.getSessionId(), event: "warnings", count: warnings.length, warnings: warnings.slice(0, 5) });
   }
   const lines = [`\u25A3 ACP | ${formatK2(beforeTokens)} \u2192 ${formatK2(afterTokens)} tokens (~${formatK2(reclaimed)} reclaimed, ${blocksCreated} block${blocksCreated > 1 ? "s" : ""})`];
+  if (blocksCreated > 0) {
+    lines.push(`Compressed: ${newBlocks.map((b2) => `${b2.blockId} (${b2.startRef ?? "?"}..${b2.endRef ?? "?"})`).join(", ")} \u2014 these refs are now consumed; do not compress them again.`);
+  }
   if (warnings.length > 0) lines.push("\u26A0\uFE0F " + warnings.join("; "));
   if (errors.length > 0) lines.push("Errors: " + errors.join("; "));
   return lines.join("\n");
@@ -12025,7 +12035,7 @@ function makeStatusTool(runtime) {
 }
 async function handleStatus(args, runtime, ctx) {
   const { state, coreMessages, entries } = await runtime.stateFor(ctx);
-  const config = applyOutputHeadroom(runtime.configFor(ctx), ctx.model);
+  const config = applyOutputHeadroom(runtime.configFor(ctx), ctx.model, resolveOutputHeadroomCap(runtime.adapter.outputHeadroomMaxPct));
   const coveredIds = collectCoveredMessageIds(state);
   const systemPromptText = getSystemPromptText(ctx);
   const systemPromptTokens = systemPromptText ? defaultCountTokens(systemPromptText) : 0;
@@ -16521,7 +16531,7 @@ ${text}`);
 }
 async function statusReport(runtime, ctx) {
   const { state, coreMessages, entries } = await runtime.stateFor(ctx);
-  const config = applyOutputHeadroom(runtime.configFor(ctx), ctx.model);
+  const config = applyOutputHeadroom(runtime.configFor(ctx), ctx.model, resolveOutputHeadroomCap(runtime.adapter.outputHeadroomMaxPct));
   const realUsage = ctx.getContextUsage?.();
   const anchorStale = usageAnchorPredatesCompression(entries ?? []);
   const systemPromptText = getSystemPromptText(ctx);
@@ -16533,7 +16543,7 @@ async function statusReport(runtime, ctx) {
   const sentTokens = estimateTokens(coreMessages, coveredIds, imageTokens) + systemPromptTokens;
   const viewSentTokens = adjustedTokenCount(runtime.core, coreMessages, state, config, sentTokens, imageTokens, systemPromptTokens);
   const turn = runtime.core.processTurn({ messages: coreMessages, state, config, tokenCount: anchorStale ? viewSentTokens : Math.max(viewSentTokens, realUsage?.tokens ?? 0) });
-  const versionStr = "0.1.59" ? `billion-context-pi@${"0.1.59"}` : void 0;
+  const versionStr = "0.1.60" ? `billion-context-pi@${"0.1.60"}` : void 0;
   let text = buildStatusPanel({
     version: versionStr,
     tokenCount: sessionTokens,
@@ -17067,7 +17077,7 @@ async function autoInstallLatest(latest, extDirOverride) {
       "--no-fund",
       "--no-save"
     ];
-    const prevVersion = (await readPackageJson(join11(extDir, "package.json")))?.version ?? "0.1.59";
+    const prevVersion = (await readPackageJson(join11(extDir, "package.json")))?.version ?? "0.1.60";
     const { code, stderr } = await runNpmImpl(installArgs(latest), { cwd: npmDir, timeout: 6e4 });
     if (code !== 0) {
       logWarn("update", {
@@ -17085,7 +17095,7 @@ async function autoInstallLatest(latest, extDirOverride) {
     }
     const verify = await verifyInstall(npmDir, latest);
     if (!verify.ok) {
-      const rollbackTo = SEMVER_RE.test(prevVersion) ? prevVersion : "0.1.59";
+      const rollbackTo = SEMVER_RE.test(prevVersion) ? prevVersion : "0.1.60";
       logWarn("update", { event: "auto-install-verify-failed", latest, reason: verify.reason, rollbackTo });
       const rb = await runNpmImpl(installArgs(rollbackTo), { cwd: npmDir, timeout: 6e4 });
       logInfo("update", { event: "rollback", from: latest, to: rollbackTo, ok: rb.code === 0 });
@@ -17160,7 +17170,7 @@ async function checkForUpdate(autoUpdate, notify) {
     }
     const latest = await fetchLatestVersion(tag);
     if (!latest) return;
-    const current = runtimeVersion ?? "0.1.59";
+    const current = runtimeVersion ?? "0.1.60";
     const hasUpdate = isVersionNewer(latest, current);
     debug.event("update-check", {
       current,
@@ -17338,7 +17348,7 @@ function wireSessionLifecycle(pi, runtime, standDownIfProxied) {
     setDelegatePolicy(DEFAULT_DELEGATE_POLICY);
     const sid = ctx.sessionManager.getSessionId();
     const modelInfo = ctx.model;
-    logInfo("session", { event: "start", sid, cwd: ctx.cwd, debug: runtime.adapter.debug ?? null, version: true ? "0.1.59" : null, model: modelInfo?.id ?? null, modelApi: modelInfo?.api ?? null, contextWindow: modelInfo?.contextWindow ?? null });
+    logInfo("session", { event: "start", sid, cwd: ctx.cwd, debug: runtime.adapter.debug ?? null, version: true ? "0.1.60" : null, model: modelInfo?.id ?? null, modelApi: modelInfo?.api ?? null, contextWindow: modelInfo?.contextWindow ?? null });
     try {
       await runtime.reloadConfig(ctx.cwd);
       const delegateCfg = resolveDelegate(runtime.adapter);
@@ -17399,10 +17409,11 @@ function wireContextTransform(pi, runtime, standDownIfProxied) {
         config = { ...config, modelContextLimit: learnedWindow };
         logInfo("overflow-selfheal", { sid, modelId, event: "window-recenter", resolved: configBase.modelContextLimit, learned: learnedWindow });
       }
-      const beforeHeadroom = config.modelContextLimit;
-      config = applyOutputHeadroom(config, ctx.model);
-      if (config.modelContextLimit !== beforeHeadroom) {
-        logInfo("overflow-selfheal", { sid, event: "output-headroom", before: beforeHeadroom, after: config.modelContextLimit, maxOutput: ctx.model?.maxTokens ?? 0 });
+      const headroomCap = resolveOutputHeadroomCap(runtime.adapter.outputHeadroomMaxPct);
+      const fullWindow = config.modelContextLimit;
+      config = applyOutputHeadroom(config, ctx.model, headroomCap);
+      if (config.modelContextLimit !== fullWindow) {
+        logInfo("overflow-selfheal", { sid, event: "output-headroom", before: fullWindow, after: config.modelContextLimit, maxOutput: ctx.model?.maxTokens ?? 0, cap: headroomCap });
       }
       const coveredIds = collectCoveredMessageIds(state);
       const realUsage = ctx.getContextUsage?.();
@@ -17456,6 +17467,7 @@ function wireContextTransform(pi, runtime, standDownIfProxied) {
         tokens: tokenCount,
         pct: config.modelContextLimit > 0 ? Number((tokenCount / config.modelContextLimit * 100).toFixed(2)) : null,
         limit: config.modelContextLimit,
+        ...fullWindow !== config.modelContextLimit ? { fullWindow } : {},
         nudge: turn.nudge?.shouldInject ? turn.nudge.breakdown?.emergencyOverride === 1 ? "emergency" : "active" : "idle",
         nudgeReason: turn.nudge?.reason ?? null,
         blocks: turn.state.blocks.length,