fix: update currentSequenceModel when modelChanged (#17051)

adamfweidman · scidomino · commit 217f27758050 · 2026-01-20T09:40:43.000-08:00
diff --git a/packages/cli/src/ui/hooks/useQuotaAndFallback.test.ts b/packages/cli/src/ui/hooks/useQuotaAndFallback.test.ts
@@ -166,11 +166,6 @@ describe('useQuotaAndFallback', () => {
         const intent = await promise!;
         expect(intent).toBe('retry_always');
 
-        // Verify activateFallbackMode was called
-        expect(mockConfig.activateFallbackMode).toHaveBeenCalledWith(
-          'gemini-flash',
-        );
-
         // The pending request should be cleared from the state
         expect(result.current.proQuotaRequest).toBeNull();
         expect(mockHistoryManager.addItem).toHaveBeenCalledTimes(1);
@@ -282,11 +277,6 @@ describe('useQuotaAndFallback', () => {
           const intent = await promise!;
           expect(intent).toBe('retry_always');
 
-          // Verify activateFallbackMode was called
-          expect(mockConfig.activateFallbackMode).toHaveBeenCalledWith(
-            'model-B',
-          );
-
           // The pending request should be cleared from the state
           expect(result.current.proQuotaRequest).toBeNull();
           expect(mockConfig.setQuotaErrorOccurred).toHaveBeenCalledWith(true);
@@ -342,11 +332,6 @@ To disable gemini-3-pro-preview, disable "Preview features" in /settings.`,
         const intent = await promise!;
         expect(intent).toBe('retry_always');
 
-        // Verify activateFallbackMode was called
-        expect(mockConfig.activateFallbackMode).toHaveBeenCalledWith(
-          'gemini-2.5-pro',
-        );
-
         expect(result.current.proQuotaRequest).toBeNull();
       });
     });
@@ -430,11 +415,6 @@ To disable gemini-3-pro-preview, disable "Preview features" in /settings.`,
       expect(intent).toBe('retry_always');
       expect(result.current.proQuotaRequest).toBeNull();
 
-      // Verify activateFallbackMode was called
-      expect(mockConfig.activateFallbackMode).toHaveBeenCalledWith(
-        'gemini-flash',
-      );
-
       // Verify quota error flags are reset
       expect(mockSetModelSwitchedFromQuotaError).toHaveBeenCalledWith(false);
       expect(mockConfig.setQuotaErrorOccurred).toHaveBeenCalledWith(false);
diff --git a/packages/cli/src/ui/hooks/useQuotaAndFallback.ts b/packages/cli/src/ui/hooks/useQuotaAndFallback.ts
@@ -135,10 +135,6 @@ export function useQuotaAndFallback({
         config.setQuotaErrorOccurred(false);
 
         if (choice === 'retry_always') {
-          // Set the model to the fallback model for the current session.
-          // This ensures the Footer updates and future turns use this model.
-          // The change is not persisted, so the original model is restored on restart.
-          config.activateFallbackMode(proQuotaRequest.fallbackModel);
           historyManager.addItem(
             {
               type: MessageType.INFO,
diff --git a/packages/core/src/config/config.ts b/packages/core/src/config/config.ts
@@ -1955,9 +1955,8 @@ export class Config {
    */
   async dispose(): Promise<void> {
     coreEvents.off(CoreEvent.AgentsRefreshed, this.onAgentsRefreshed);
-    if (this.agentRegistry) {
-      this.agentRegistry.dispose();
-    }
+    this.agentRegistry?.dispose();
+    this.geminiClient?.dispose();
     if (this.mcpClientManager) {
       await this.mcpClientManager.stop();
     }
diff --git a/packages/core/src/core/client.test.ts b/packages/core/src/core/client.test.ts
@@ -48,6 +48,7 @@ import type {
 import { ClearcutLogger } from '../telemetry/clearcut-logger/clearcut-logger.js';
 import * as policyCatalog from '../availability/policyCatalog.js';
 import { partToString } from '../utils/partUtils.js';
+import { coreEvents } from '../utils/events.js';
 
 vi.mock('../services/chatCompressionService.js');
 
@@ -290,6 +291,7 @@ describe('Gemini Client (client.ts)', () => {
   });
 
   afterEach(() => {
+    client.dispose();
     vi.restoreAllMocks();
   });
 
@@ -1579,6 +1581,55 @@ ${JSON.stringify(
           expect.any(AbortSignal),
         );
       });
+
+      it('should re-route within the same prompt when the configured model changes', async () => {
+        mockTurnRunFn.mockClear();
+        mockTurnRunFn.mockImplementation(async function* () {
+          yield { type: 'content', value: 'Hello' };
+        });
+
+        mockRouterService.route.mockResolvedValueOnce({
+          model: 'original-model',
+          reason: 'test',
+        });
+
+        let stream = client.sendMessageStream(
+          [{ text: 'Hi' }],
+          new AbortController().signal,
+          'prompt-1',
+        );
+        await fromAsync(stream);
+
+        expect(mockRouterService.route).toHaveBeenCalledTimes(1);
+        expect(mockTurnRunFn).toHaveBeenNthCalledWith(
+          1,
+          { model: 'original-model' },
+          [{ text: 'Hi' }],
+          expect.any(AbortSignal),
+        );
+
+        mockRouterService.route.mockResolvedValue({
+          model: 'fallback-model',
+          reason: 'test',
+        });
+        vi.mocked(mockConfig.getModel).mockReturnValue('gemini-2.5-flash');
+        coreEvents.emitModelChanged('gemini-2.5-flash');
+
+        stream = client.sendMessageStream(
+          [{ text: 'Continue' }],
+          new AbortController().signal,
+          'prompt-1',
+        );
+        await fromAsync(stream);
+
+        expect(mockRouterService.route).toHaveBeenCalledTimes(2);
+        expect(mockTurnRunFn).toHaveBeenNthCalledWith(
+          2,
+          { model: 'fallback-model' },
+          [{ text: 'Continue' }],
+          expect.any(AbortSignal),
+        );
+      });
     });
 
     it('should use getGlobalMemory for system instruction when JIT is enabled', async () => {
diff --git a/packages/core/src/core/client.ts b/packages/core/src/core/client.ts
@@ -58,6 +58,7 @@ import {
 import { resolveModel } from '../config/models.js';
 import type { RetryAvailabilityContext } from '../utils/retry.js';
 import { partToString } from '../utils/partUtils.js';
+import { coreEvents, CoreEvent } from '../utils/events.js';
 
 const MAX_TURNS = 100;
 
@@ -94,8 +95,14 @@ export class GeminiClient {
     this.loopDetector = new LoopDetectionService(config);
     this.compressionService = new ChatCompressionService();
     this.lastPromptId = this.config.getSessionId();
+
+    coreEvents.on(CoreEvent.ModelChanged, this.handleModelChanged);
   }
 
+  private handleModelChanged = () => {
+    this.currentSequenceModel = null;
+  };
+
   // Hook state to deduplicate BeforeAgent calls and track response for
   // AfterAgent
   private hookStateMap = new Map<
@@ -253,6 +260,10 @@ export class GeminiClient {
     this.updateTelemetryTokenCount();
   }
 
+  dispose() {
+    coreEvents.off(CoreEvent.ModelChanged, this.handleModelChanged);
+  }
+
   async resumeChat(
     history: Content[],
     resumedSessionData?: ResumedSessionData,
diff --git a/packages/core/src/fallback/handler.test.ts b/packages/core/src/fallback/handler.test.ts
@@ -65,6 +65,8 @@ const createMockConfig = (overrides: Partial<Config> = {}): Config =>
     fallbackHandler: undefined,
     getFallbackModelHandler: vi.fn(),
     setActiveModel: vi.fn(),
+    setModel: vi.fn(),
+    activateFallbackMode: vi.fn(),
     getModelAvailabilityService: vi.fn(() =>
       createAvailabilityServiceMock({
         selectedModel: FALLBACK_MODEL,
@@ -198,7 +200,7 @@ describe('handleFallback', () => {
 
         expect(result).toBe(true);
         expect(policyConfig.getFallbackModelHandler).not.toHaveBeenCalled();
-        expect(policyConfig.setActiveModel).toHaveBeenCalledWith(
+        expect(policyConfig.activateFallbackMode).toHaveBeenCalledWith(
           DEFAULT_GEMINI_FLASH_MODEL,
         );
       } finally {
@@ -273,7 +275,7 @@ describe('handleFallback', () => {
       expect(openBrowserSecurely).toHaveBeenCalledWith(
         'https://goo.gle/set-up-gemini-code-assist',
       );
-      expect(policyConfig.setActiveModel).not.toHaveBeenCalled();
+      expect(policyConfig.activateFallbackMode).not.toHaveBeenCalled();
     });
 
     it('should catch errors from the handler, log an error, and return null', async () => {
@@ -378,7 +380,7 @@ describe('handleFallback', () => {
       );
     });
 
-    it('calls setActiveModel and logs telemetry when handler returns "retry_always"', async () => {
+    it('calls activateFallbackMode when handler returns "retry_always"', async () => {
       policyHandler.mockResolvedValue('retry_always');
       vi.mocked(policyConfig.getModel).mockReturnValue(
         DEFAULT_GEMINI_MODEL_AUTO,
@@ -391,11 +393,13 @@ describe('handleFallback', () => {
       );
 
       expect(result).toBe(true);
-      expect(policyConfig.setActiveModel).toHaveBeenCalledWith(FALLBACK_MODEL);
+      expect(policyConfig.activateFallbackMode).toHaveBeenCalledWith(
+        FALLBACK_MODEL,
+      );
       // TODO: add logging expect statement
     });
 
-    it('does NOT call setActiveModel when handler returns "stop"', async () => {
+    it('does NOT call activateFallbackMode when handler returns "stop"', async () => {
       policyHandler.mockResolvedValue('stop');
 
       const result = await handleFallback(
@@ -405,11 +409,11 @@ describe('handleFallback', () => {
       );
 
       expect(result).toBe(false);
-      expect(policyConfig.setActiveModel).not.toHaveBeenCalled();
+      expect(policyConfig.activateFallbackMode).not.toHaveBeenCalled();
       // TODO: add logging expect statement
     });
 
-    it('does NOT call setActiveModel when handler returns "retry_once"', async () => {
+    it('does NOT call activateFallbackMode when handler returns "retry_once"', async () => {
       policyHandler.mockResolvedValue('retry_once');
 
       const result = await handleFallback(
@@ -419,7 +423,7 @@ describe('handleFallback', () => {
       );
 
       expect(result).toBe(true);
-      expect(policyConfig.setActiveModel).not.toHaveBeenCalled();
+      expect(policyConfig.activateFallbackMode).not.toHaveBeenCalled();
     });
   });
 });
diff --git a/packages/core/src/fallback/handler.ts b/packages/core/src/fallback/handler.ts
@@ -131,7 +131,7 @@ async function processIntent(
     case 'retry_always':
       // TODO(telemetry): Implement generic fallback event logging. Existing
       // logFlashFallback is specific to a single Model.
-      config.setActiveModel(fallbackModel);
+      config.activateFallbackMode(fallbackModel);
       return true;
 
     case 'retry_once':
