@google/gemini-cli-a2a-server 0.11.2 → 0.11.3

package/dist/a2a-server.mjs

@@ -288503,8 +288503,8 @@ var Float64Vector = import_vector.default.Float64Vector;
 var PointerVector = import_vector.default.PointerVector;
 
 // packages/core/dist/src/generated/git-commit.js
-var GIT_COMMIT_INFO = "f36dec6a";
-var CLI_VERSION = "0.11.2";
+var GIT_COMMIT_INFO = "44b3c974";
+var CLI_VERSION = "0.11.3";
 
 // packages/core/dist/src/ide/detect-ide.js
 var IDE_DEFINITIONS = {
@@ -290970,7 +290970,7 @@ function createContentGeneratorConfig(config2, authType) {
   return contentGeneratorConfig;
 }
 async function createContentGenerator(config2, gcConfig, sessionId2) {
-  const version3 = "0.11.2";
+  const version3 = "0.11.3";
   const userAgent = `GeminiCLI/${version3} (${process.platform}; ${process.arch})`;
   const baseHeaders = {
     "User-Agent": userAgent
@@ -324782,8 +324782,7 @@ function classifyGoogleError(error) {
     for (const violation of quotaFailure.violations) {
       const quotaId = violation.quotaId ?? "";
       if (quotaId.includes("PerDay") || quotaId.includes("Daily")) {
-        return new TerminalQuotaError(`${googleApiError.message}
-Expected quota reset within 24h.`, googleApiError);
+        return new TerminalQuotaError(`You have exhausted your daily quota on this model.`, googleApiError);
       }
     }
   }
@@ -324812,8 +324811,7 @@ Expected quota reset within 24h.`, googleApiError);
     }
     const quotaLimit = errorInfo.metadata?.["quota_limit"] ?? "";
     if (quotaLimit.includes("PerDay") || quotaLimit.includes("Daily")) {
-      return new TerminalQuotaError(`${googleApiError.message}
-Expected quota reset within 24h.`, googleApiError);
+      return new TerminalQuotaError(`You have exhausted your daily quota on this model.`, googleApiError);
     }
   }
   if (retryInfo?.retryDelay) {
@@ -337928,6 +337926,24 @@ ${directoryContext}
   const initialParts = [{ text: context2 }];
   return initialParts;
 }
+async function getInitialChatHistory(config2, extraHistory) {
+  const envParts = await getEnvironmentContext(config2);
+  const envContextString = envParts.map((part) => part.text || "").join("\n\n");
+  const allSetupText = `
+${envContextString}
+
+Reminder: Do not return an empty response when a tool call is required.
+
+My setup is complete. I will provide my first command in the next turn.
+    `.trim();
+  return [
+    {
+      role: "user",
+      parts: [{ text: allSetupText }]
+    },
+    ...extraHistory ?? []
+  ];
+}
 
 // packages/core/dist/src/utils/generateContentResponseUtilities.js
 function getResponseTextFromParts(parts2) {
@@ -340034,7 +340050,10 @@ var LoopDetectionService = class {
    * @returns true if a loop is detected, false otherwise
    */
   addAndCheck(event) {
-    if (this.loopDetected || this.disabledForSession) {
+    if (this.disabledForSession) {
+      return false;
+    }
+    if (this.loopDetected) {
       return this.loopDetected;
     }
     switch (event.type) {
@@ -340282,10 +340301,9 @@ var LoopDetectionService = class {
   }
 };
 
-// packages/core/dist/src/core/client.js
-function isThinkingSupported(model) {
-  return model.startsWith("gemini-2.5") || model === DEFAULT_GEMINI_MODEL_AUTO;
-}
+// packages/core/dist/src/services/chatCompressionService.js
+var DEFAULT_COMPRESSION_TOKEN_THRESHOLD = 0.2;
+var COMPRESSION_PRESERVE_THRESHOLD = 0.3;
 function findCompressSplitPoint(contents, fraction) {
   if (fraction <= 0 || fraction >= 1) {
     throw new Error("Fraction must be between 0 and 1");
@@ -340311,9 +340329,109 @@ function findCompressSplitPoint(contents, fraction) {
   }
   return lastSplitPoint;
 }
+var ChatCompressionService = class {
+  async compress(chat, promptId, force, model, config2, hasFailedCompressionAttempt) {
+    const curatedHistory = chat.getHistory(true);
+    if (curatedHistory.length === 0 || hasFailedCompressionAttempt && !force) {
+      return {
+        newHistory: null,
+        info: {
+          originalTokenCount: 0,
+          newTokenCount: 0,
+          compressionStatus: CompressionStatus.NOOP
+        }
+      };
+    }
+    const originalTokenCount = uiTelemetryService.getLastPromptTokenCount();
+    if (!force) {
+      const threshold = config2.getCompressionThreshold() ?? DEFAULT_COMPRESSION_TOKEN_THRESHOLD;
+      if (originalTokenCount < threshold * tokenLimit(model)) {
+        return {
+          newHistory: null,
+          info: {
+            originalTokenCount,
+            newTokenCount: originalTokenCount,
+            compressionStatus: CompressionStatus.NOOP
+          }
+        };
+      }
+    }
+    const splitPoint = findCompressSplitPoint(curatedHistory, 1 - COMPRESSION_PRESERVE_THRESHOLD);
+    const historyToCompress = curatedHistory.slice(0, splitPoint);
+    const historyToKeep = curatedHistory.slice(splitPoint);
+    if (historyToCompress.length === 0) {
+      return {
+        newHistory: null,
+        info: {
+          originalTokenCount,
+          newTokenCount: originalTokenCount,
+          compressionStatus: CompressionStatus.NOOP
+        }
+      };
+    }
+    const summaryResponse = await config2.getContentGenerator().generateContent({
+      model,
+      contents: [
+        ...historyToCompress,
+        {
+          role: "user",
+          parts: [
+            {
+              text: "First, reason in your scratchpad. Then, generate the <state_snapshot>."
+            }
+          ]
+        }
+      ],
+      config: {
+        systemInstruction: { text: getCompressionPrompt() }
+      }
+    }, promptId);
+    const summary = getResponseText(summaryResponse) ?? "";
+    const extraHistory = [
+      {
+        role: "user",
+        parts: [{ text: summary }]
+      },
+      {
+        role: "model",
+        parts: [{ text: "Got it. Thanks for the additional context!" }]
+      },
+      ...historyToKeep
+    ];
+    const fullNewHistory = await getInitialChatHistory(config2, extraHistory);
+    const newTokenCount = Math.floor(fullNewHistory.reduce((total, content) => total + JSON.stringify(content).length, 0) / 4);
+    logChatCompression(config2, makeChatCompressionEvent({
+      tokens_before: originalTokenCount,
+      tokens_after: newTokenCount
+    }));
+    if (newTokenCount > originalTokenCount) {
+      return {
+        newHistory: null,
+        info: {
+          originalTokenCount,
+          newTokenCount,
+          compressionStatus: CompressionStatus.COMPRESSION_FAILED_INFLATED_TOKEN_COUNT
+        }
+      };
+    } else {
+      uiTelemetryService.setLastPromptTokenCount(newTokenCount);
+      return {
+        newHistory: extraHistory,
+        info: {
+          originalTokenCount,
+          newTokenCount,
+          compressionStatus: CompressionStatus.COMPRESSED
+        }
+      };
+    }
+  }
+};
+
+// packages/core/dist/src/core/client.js
+function isThinkingSupported(model) {
+  return model.startsWith("gemini-2.5") || model === DEFAULT_GEMINI_MODEL_AUTO;
+}
 var MAX_TURNS = 100;
-var COMPRESSION_TOKEN_THRESHOLD = 0.7;
-var COMPRESSION_PRESERVE_THRESHOLD = 0.3;
 var GeminiClient = class {
   config;
   chat;
@@ -340323,6 +340441,7 @@ var GeminiClient = class {
   };
   sessionTurnCount = 0;
   loopDetector;
+  compressionService;
   lastPromptId;
   currentSequenceModel = null;
   lastSentIdeContext;
@@ -340335,6 +340454,7 @@ var GeminiClient = class {
   constructor(config2) {
     this.config = config2;
     this.loopDetector = new LoopDetectionService(config2);
+    this.compressionService = new ChatCompressionService();
     this.lastPromptId = this.config.getSessionId();
   }
   async initialize() {
@@ -340401,22 +340521,7 @@ var GeminiClient = class {
     const toolRegistry = this.config.getToolRegistry();
     const toolDeclarations = toolRegistry.getFunctionDeclarations();
     const tools = [{ functionDeclarations: toolDeclarations }];
-    const envParts = await getEnvironmentContext(this.config);
-    const envContextString = envParts.map((part) => part.text || "").join("\n\n");
-    const allSetupText = `
-${envContextString}
-
-Reminder: Do not return an empty response when a tool call is required.
-
-My setup is complete. I will provide my first command in the next turn.
-    `.trim();
-    const history = [
-      {
-        role: "user",
-        parts: [{ text: allSetupText }]
-      },
-      ...extraHistory ?? []
-    ];
+    const history = await getInitialChatHistory(this.config, extraHistory);
     try {
       const userMemory = this.config.getUserMemory();
       const systemInstruction = getCoreSystemPrompt(this.config, userMemory);
@@ -340728,87 +340833,16 @@ My setup is complete. I will provide my first command in the next turn.
   }
   async tryCompressChat(prompt_id, force = false) {
     const model = this._getEffectiveModelForCurrentTurn();
-    const curatedHistory = this.getChat().getHistory(true);
-    if (curatedHistory.length === 0 || this.hasFailedCompressionAttempt && !force) {
-      return {
-        originalTokenCount: 0,
-        newTokenCount: 0,
-        compressionStatus: CompressionStatus.NOOP
-      };
-    }
-    const originalTokenCount = uiTelemetryService.getLastPromptTokenCount();
-    const contextPercentageThreshold = this.config.getChatCompression()?.contextPercentageThreshold;
-    if (!force) {
-      const threshold = contextPercentageThreshold ?? COMPRESSION_TOKEN_THRESHOLD;
-      if (originalTokenCount < threshold * tokenLimit(model)) {
-        return {
-          originalTokenCount,
-          newTokenCount: originalTokenCount,
-          compressionStatus: CompressionStatus.NOOP
-        };
-      }
-    }
-    const splitPoint = findCompressSplitPoint(curatedHistory, 1 - COMPRESSION_PRESERVE_THRESHOLD);
-    const historyToCompress = curatedHistory.slice(0, splitPoint);
-    const historyToKeep = curatedHistory.slice(splitPoint);
-    if (historyToCompress.length === 0) {
-      return {
-        originalTokenCount,
-        newTokenCount: originalTokenCount,
-        compressionStatus: CompressionStatus.NOOP
-      };
-    }
-    const summaryResponse = await this.config.getContentGenerator().generateContent({
-      model,
-      contents: [
-        ...historyToCompress,
-        {
-          role: "user",
-          parts: [
-            {
-              text: "First, reason in your scratchpad. Then, generate the <state_snapshot>."
-            }
-          ]
-        }
-      ],
-      config: {
-        systemInstruction: { text: getCompressionPrompt() }
-      }
-    }, prompt_id);
-    const summary = getResponseText(summaryResponse) ?? "";
-    const chat = await this.startChat([
-      {
-        role: "user",
-        parts: [{ text: summary }]
-      },
-      {
-        role: "model",
-        parts: [{ text: "Got it. Thanks for the additional context!" }]
-      },
-      ...historyToKeep
-    ]);
-    this.forceFullIdeContext = true;
-    const newTokenCount = Math.floor(chat.getHistory().reduce((total, content) => total + JSON.stringify(content).length, 0) / 4);
-    logChatCompression(this.config, makeChatCompressionEvent({
-      tokens_before: originalTokenCount,
-      tokens_after: newTokenCount
-    }));
-    if (newTokenCount > originalTokenCount) {
+    const { newHistory, info: info2 } = await this.compressionService.compress(this.getChat(), prompt_id, force, model, this.config, this.hasFailedCompressionAttempt);
+    if (info2.compressionStatus === CompressionStatus.COMPRESSION_FAILED_INFLATED_TOKEN_COUNT) {
       this.hasFailedCompressionAttempt = !force && true;
-      return {
-        originalTokenCount,
-        newTokenCount,
-        compressionStatus: CompressionStatus.COMPRESSION_FAILED_INFLATED_TOKEN_COUNT
-      };
-    } else {
-      this.chat = chat;
-      uiTelemetryService.setLastPromptTokenCount(newTokenCount);
+    } else if (info2.compressionStatus === CompressionStatus.COMPRESSED) {
+      if (newHistory) {
+        this.chat = await this.startChat(newHistory);
+        this.forceFullIdeContext = true;
+      }
     }
-    return {
-      originalTokenCount,
-      newTokenCount,
-      compressionStatus: CompressionStatus.COMPRESSED
-    };
+    return info2;
   }
 };
 
@@ -348676,7 +348710,7 @@ var Config = class {
   summarizeToolOutput;
   experimentalZedIntegration = false;
   loadMemoryFromIncludeDirectories = false;
-  chatCompression;
+  compressionThreshold;
   interactive;
   ptyInfo;
   trustedFolder;
@@ -348761,7 +348795,7 @@ var Config = class {
     this.folderTrust = params.folderTrust ?? false;
     this.ideMode = params.ideMode ?? false;
     this.loadMemoryFromIncludeDirectories = params.loadMemoryFromIncludeDirectories ?? false;
-    this.chatCompression = params.chatCompression;
+    this.compressionThreshold = params.compressionThreshold;
     this.interactive = params.interactive ?? false;
     this.ptyInfo = params.ptyInfo ?? "child_process";
     this.trustedFolder = params.trustedFolder;
@@ -349135,8 +349169,8 @@ var Config = class {
   setFileSystemService(fileSystemService) {
     this.fileSystemService = fileSystemService;
   }
-  getChatCompression() {
-    return this.chatCompression;
+  getCompressionThreshold() {
+    return this.compressionThreshold;
   }
   isInteractiveShellEnabled() {
     return this.interactive && this.ptyInfo !== "child_process" && this.enableInteractiveShell;