Compare commits

...

1 Commits

Author SHA1 Message Date
易良
cea15a118b
fix(core,cli): replace full-history structuredClone with shallow/tail variants to prevent OOM on resume (#4644)
* fix(core,cli): replace full-history structuredClone with shallow/tail variants to prevent OOM on resume

Several UI and service call sites clone the entire chat history via
structuredClone(getHistory()) every turn. On a resumed session with
thousands of entries, each clone allocates 150-200 MB transiently.
When multiple async side-requests overlap (suggestion generation,
auto-title, checkpointing), multiple clones coexist on the heap,
pushing V8 past its limit within 10 turns (2 GB heap cap).

Changes:
- AppContainer.tsx: use getHistoryTail(40, true) instead of
  getHistory(true) + slice(-40)
- btwCommand.ts: same pattern, use getHistoryTail(40, true)
- sessionTitle.ts: use getHistoryShallow() (read-only filtering)
- sessionRecap.ts: use getHistoryShallow() (read-only filtering)
- useGeminiStream.ts: use getHistoryShallow() for checkpoint
  serialization (only needs to survive JSON.stringify)

Closes #4624

* fix(test): update mocks for getHistoryShallow/getHistoryTail in sessionTitle and btwCommand tests

* fix(cli): migrate remaining getHistory() clone sites to shallow/tail variants

- AppContainer.tsx rewind path: getHistory() → getHistoryShallow()
  (only used read-only by computeApiTruncationIndex)
- Session.ts ACP rewind: getHistory() → getHistoryShallow()
  (only walks entries to compute truncation index)
- Session.ts stop-hook: getHistory() + filter(.model).pop() →
  getLastModelMessageText() (O(1) backward scan, no clone)

* fix(core): use client-level getHistoryShallow with fallback

sessionTitle.ts and sessionRecap.ts were calling
chat.getHistoryShallow() directly, bypassing the client-level
wrapper that provides a getHistory() fallback when the chat
implementation doesn't support shallow reads. Use
geminiClient.getHistoryShallow() instead.

Update test mocks to match the new call site.

* fix(test): add getHistoryShallow and getLastModelMessageText to Session test mocks

Session.ts now calls chat.getHistoryShallow() in rewindToTurn and
chat.getLastModelMessageText() in the Stop hook. Update all mockChat
instances in Session.test.ts to provide these methods.
2026-06-01 22:53:48 +08:00
9 changed files with 60 additions and 30 deletions

View File

@ -196,6 +196,8 @@ describe('Session', () => {
sendMessageStream: vi.fn(),
addHistory: vi.fn(),
getHistory: vi.fn().mockReturnValue([]),
getHistoryShallow: vi.fn().mockReturnValue([]),
getLastModelMessageText: vi.fn().mockReturnValue(''),
setHistory: vi.fn(),
truncateHistory: vi.fn(),
stripThoughtsFromHistory: vi.fn(),
@ -329,6 +331,7 @@ describe('Session', () => {
{ role: 'model', parts: [{ text: 'second reply' }] },
];
vi.mocked(mockChat.getHistory).mockReturnValue(history);
vi.mocked(mockChat.getHistoryShallow).mockReturnValue(history);
const result = session.rewindToTurn(1);
@ -348,6 +351,7 @@ describe('Session', () => {
{ role: 'model', parts: [{ text: 'first reply' }] },
];
vi.mocked(mockChat.getHistory).mockReturnValue(history);
vi.mocked(mockChat.getHistoryShallow).mockReturnValue(history);
const result = session.rewindToTurn(0);
@ -356,9 +360,9 @@ describe('Session', () => {
});
it('rejects unreachable user turns', () => {
vi.mocked(mockChat.getHistory).mockReturnValue([
{ role: 'user', parts: [{ text: 'first' }] },
]);
const history: Content[] = [{ role: 'user', parts: [{ text: 'first' }] }];
vi.mocked(mockChat.getHistory).mockReturnValue(history);
vi.mocked(mockChat.getHistoryShallow).mockReturnValue(history);
expect(() => session.rewindToTurn(2)).toThrow(
'Cannot rewind to the requested turn',
@ -853,6 +857,8 @@ describe('Session', () => {
sendMessageStream: vi.fn().mockResolvedValue(createEmptyStream()),
addHistory: vi.fn(),
getHistory: vi.fn().mockReturnValue([]),
getHistoryShallow: vi.fn().mockReturnValue([]),
getLastModelMessageText: vi.fn().mockReturnValue(''),
} as unknown as GeminiChat;
mockChat.sendMessageStream = vi
@ -1207,6 +1213,8 @@ describe('Session', () => {
sendMessageStream: vi.fn().mockResolvedValue(createEmptyStream()),
addHistory: vi.fn(),
getHistory: vi.fn().mockReturnValue([]),
getHistoryShallow: vi.fn().mockReturnValue([]),
getLastModelMessageText: vi.fn().mockReturnValue(''),
} as unknown as GeminiChat;
mockConfig.getSessionTokenLimit = vi.fn().mockReturnValue(100);
mockGeminiClient.tryCompressChat
@ -1523,6 +1531,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi
.fn()
.mockResolvedValueOnce(createEmptyStream())
@ -1584,6 +1595,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi
.fn()
.mockResolvedValueOnce(createEmptyStream())
@ -1655,6 +1669,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi
.fn()
.mockResolvedValue(createEmptyStream());
@ -2405,6 +2422,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi.fn().mockResolvedValue(
createStreamWithChunks([
@ -2458,6 +2478,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi
.fn()
.mockResolvedValue(createEmptyStream());
@ -2503,6 +2526,9 @@ describe('Session', () => {
.mockReturnValue([
{ role: 'model', parts: [{ text: 'response text' }] },
]);
mockChat.getLastModelMessageText = vi
.fn()
.mockReturnValue('response text');
mockChat.sendMessageStream = vi
.fn()
.mockResolvedValue(createEmptyStream());

View File

@ -402,7 +402,7 @@ export class Session implements SessionContext {
}
const chat = this.config.getGeminiClient()!.getChat();
const apiHistory = chat.getHistory();
const apiHistory = chat.getHistoryShallow();
const apiTruncateIndex = this.#computeApiTruncationIndexForUserTurn(
apiHistory,
targetTurnIndex,
@ -895,16 +895,10 @@ export class Session implements SessionContext {
return { stopReason: 'end_turn' };
}
// Get response text from the chat history
const history = this.#getCurrentChat().getHistory();
const lastModelMessage = history
.filter((msg: Content) => msg.role === 'model')
.pop();
// Extract last model text without cloning the full history.
const responseText =
lastModelMessage?.parts
?.filter((p: Part): p is { text: string } & Part => 'text' in p)
.map((p: { text: string }) => p.text)
.join('') || '[no response text]';
this.#getCurrentChat().getLastModelMessageText?.() ||
'[no response text]';
const response = await messageBus.request<
HookExecutionRequest,

View File

@ -2108,10 +2108,9 @@ export const AppContainer = (props: AppContainerProps) => {
const ac = new AbortController();
suggestionAbortRef.current = ac;
// Use curated history to avoid invalid/empty entries causing API errors
const fullHistory = geminiClient.getChat().getHistory(true);
const conversationHistory =
fullHistory.length > 40 ? fullHistory.slice(-40) : fullHistory;
// Only clone the tail — full structuredClone of a large resumed session
// causes transient heap peaks that trigger OOM (#4624).
const conversationHistory = geminiClient.getHistoryTail(40, true);
generatePromptSuggestion(config, conversationHistory, ac.signal, {
enableCacheSharing: settings.merged.ui?.enableCacheSharing === true,
})
@ -2464,7 +2463,7 @@ export const AppContainer = (props: AppContainerProps) => {
apiTruncateIndex = computeApiTruncationIndex(
historyManager.history,
userItem.id,
geminiClient.getHistory(),
geminiClient.getHistoryShallow(),
);
if (apiTruncateIndex < 0) {
historyManager.addItem(

View File

@ -176,6 +176,11 @@ describe('btwCommand', () => {
.mockReturnValue([
{ role: 'user', parts: [{ text: '杭州天气如何?' }] },
]),
getHistoryTail: vi
.fn()
.mockReturnValue([
{ role: 'user', parts: [{ text: '杭州天气如何?' }] },
]),
getChat: vi.fn().mockReturnValue({
getGenerationConfig: vi.fn().mockReturnValue({
systemInstruction: 'You are helpful',
@ -229,6 +234,10 @@ describe('btwCommand', () => {
{ role: 'user', parts: [{ text: '杭州天气如何?' }] },
{ role: 'user', parts: [{ text: '请顺便解释一下湿度怎么看' }] },
]),
getHistoryTail: vi.fn().mockReturnValue([
{ role: 'user', parts: [{ text: '杭州天气如何?' }] },
{ role: 'user', parts: [{ text: '请顺便解释一下湿度怎么看' }] },
]),
getChat: vi.fn().mockReturnValue({
getGenerationConfig: vi.fn().mockReturnValue({
systemInstruction: 'live system prompt',

View File

@ -55,7 +55,7 @@ function getBtwCacheSafeParams(
geminiClient &&
typeof geminiClient === 'object' &&
typeof geminiClient.getChat === 'function' &&
typeof geminiClient.getHistory === 'function'
typeof geminiClient.getHistoryTail === 'function'
) {
const chat = geminiClient.getChat();
if (
@ -65,12 +65,7 @@ function getBtwCacheSafeParams(
) {
const generationConfig = chat.getGenerationConfig();
if (generationConfig) {
const fullHistory = geminiClient.getHistory(true);
const maxHistoryEntries = 40;
const history =
fullHistory.length > maxHistoryEntries
? fullHistory.slice(-maxHistoryEntries)
: fullHistory;
const history = geminiClient.getHistoryTail(40, true);
return {
generationConfig,

View File

@ -2462,7 +2462,7 @@ export const useGeminiStream = (
const toolName = toolCall.request.name;
const fileName = path.basename(filePath);
const toolCallWithSnapshotFileName = `${timestamp}-${fileName}-${toolName}.json`;
const clientHistory = await geminiClient?.getHistory();
const clientHistory = geminiClient?.getHistoryShallow();
const toolCallWithSnapshotFilePath = path.join(
checkpointDir,
toolCallWithSnapshotFileName,

View File

@ -49,7 +49,7 @@ export async function generateSessionRecap(
const geminiClient = config.getGeminiClient();
if (!geminiClient) return null;
const fullHistory = geminiClient.getChat().getHistory();
const fullHistory = geminiClient.getHistoryShallow();
if (fullHistory.length < 2) return null;
const dialog = filterToDialog(fullHistory);

View File

@ -31,6 +31,7 @@ function makeConfig(opts: MockOptions): {
getFastModel: vi.fn(() => opts.fastModel ?? undefined),
getModel: vi.fn(() => 'qwen-plus'),
getGeminiClient: vi.fn(() => ({
getHistoryShallow: () => opts.history ?? [],
getChat: () => ({
getHistory: () => opts.history ?? [],
}),
@ -202,7 +203,10 @@ describe('tryGenerateSessionTitle', () => {
getFastModel: vi.fn(() => 'qwen-turbo'),
getModel: vi.fn(() => 'qwen-plus'),
getGeminiClient: vi.fn(() => ({
getChat: () => ({ getHistory: () => history }),
getHistoryShallow: () => history,
getChat: () => ({
getHistory: () => history,
}),
})),
getBaseLlmClient: vi.fn(() => ({ generateJson })),
} as unknown as Config;
@ -240,7 +244,10 @@ describe('tryGenerateSessionTitle', () => {
getFastModel: vi.fn(() => 'qwen-turbo'),
getModel: vi.fn(() => 'qwen-plus'),
getGeminiClient: vi.fn(() => ({
getChat: () => ({ getHistory: () => history }),
getHistoryShallow: () => history,
getChat: () => ({
getHistory: () => history,
}),
})),
getBaseLlmClient: vi.fn(() => ({ generateJson })),
} as unknown as Config;

View File

@ -113,7 +113,7 @@ export async function tryGenerateSessionTitle(
const geminiClient = config.getGeminiClient();
if (!geminiClient) return { ok: false, reason: 'no_client' };
const fullHistory = geminiClient.getChat().getHistory();
const fullHistory = geminiClient.getHistoryShallow();
if (fullHistory.length < 2) return { ok: false, reason: 'empty_history' };
const dialog = filterToDialog(fullHistory);