@earendil-works/pi-coding-agent 0.79.10 → 0.80.2

package/dist/core/agent-session.js

@@ -14,7 +14,7 @@
  */
 import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs";
 import { basename, dirname } from "node:path";
-import { clampThinkingLevel, cleanupSessionResources, getSupportedThinkingLevels, isContextOverflow, modelsAreEqual, resetApiProviders, streamSimple, } from "@earendil-works/pi-ai";
+import { clampThinkingLevel, cleanupSessionResources, getSupportedThinkingLevels, isContextOverflow, modelsAreEqual, resetApiProviders, streamSimple, } from "@earendil-works/pi-ai/compat";
 import { getThemeByName, theme } from "../modes/interactive/theme/theme.js";
 import { stripFrontmatter } from "../utils/frontmatter.js";
 import { resolvePath } from "../utils/paths.js";
@@ -1459,10 +1459,12 @@ export class AgentSession {
             return await this._runAutoCompaction("overflow", willRetry);
         }
         // Case 2: Threshold - context is getting large
-        // For error messages (no usage data), estimate from last successful response.
-        // This ensures sessions that hit persistent API errors (e.g. 529) can still compact.
+        // For error messages or all-zero usage messages, estimate from the last valid response.
+        // This ensures sessions that hit persistent API errors (e.g. 529) or malformed zero-usage
+        // responses can still compact and do not reset context accounting.
         let contextTokens;
-        if (assistantMessage.stopReason === "error") {
+        const directContextTokens = assistantMessage.usage ? calculateContextTokens(assistantMessage.usage) : 0;
+        if (assistantMessage.stopReason === "error" || directContextTokens === 0) {
             const messages = this.agent.state.messages;
             const estimate = estimateContextTokens(messages);
             if (estimate.lastUsageIndex === null)
@@ -1479,7 +1481,7 @@ export class AgentSession {
             contextTokens = estimate.tokens;
         }
         else {
-            contextTokens = calculateContextTokens(assistantMessage.usage);
+            contextTokens = directContextTokens;
         }
         if (shouldCompact(contextTokens, contextWindow, settings)) {
             return await this._runAutoCompaction("threshold", false);
@@ -2424,8 +2426,8 @@ export class AgentSession {
                         const contextTokens = calculateContextTokens(assistant.usage);
                         if (contextTokens > 0) {
                             hasPostCompactionUsage = true;
+                            break;
                         }
-                        break;
                     }
                 }
             }